Patent Application
20230381668
The present invention relates generally to the field of skill training and assessment. More specifically, embodiments of the present invention relate to skill training and assessment in video games.
Performance metrics for assessing motor skill performance change over time, or across players which may be useful in gaming, physical or cognitive therapy, and other fields, exist, but are not sufficiently efficient and accurate. Contrary to traditional sports, there are few performance metrics in video games and esports (e.g., competitive multiplayer video games) for assessing a player's skill and ability. First-person shooter (FPS) video games are centered on gun and other weapon-based combat in a first-person perspective, with the player controlling a player character. One measured skill relevant to video games is flicking skill, which measures players' skill with flick shots, a standard user input to some video games, where a user may for example quickly react to a target. Flicking, rapidly moving the crosshair to an object or enemy and firing a shot to damage or destroy the target, is a fundamental facet of FPS play. A flick shot may include taking shots where a player's aim or crosshairs begin a distance away from the target, and the player quickly flicks (e.g. by controlling a video game controller or by moving a computer mouse) so the aim moves to the target, e.g. almost concurrently with the player pulling the trigger or making a shot.
Performance in FPS video games relies on acquired perceptual and motor skills. Competitive players dedicate years advancing these skills prior to playing at a professional level. Despite the rapidly growing popularity of competitive esports, the field continues to lack established objective benchmarks for individual skill. Many of the commonly used performance metrics in esports are unreliable measures of individual skill. Particularly in team games, key benchmarks (e.g., kill-death ratio, damage dealt, win-rate) confound the ability of an individual player with that of their teammates, the opposing team, or the coordination of the team as a whole. For instance, a mediocre player is able to gain a high ranking by playing with individuals of a much higher skill level, and the opposite is true for a great player with lesser-skilled teammates. Reliable individual skill assessments with intra- and inter-individual comparisons both at a single time point and longitudinally are needed to characterize player rankings and efficacy of training.
Successful FPS play requires efficient identification and localization of relevant visual stimuli, and dynamic movements followed by well-timed shot responses. When playing with a computer mouse and keyboard, for example, flicking is achieved primarily by hand and arm (e.g., reach) movements. Protocols for measuring ballistic reach and eye movement kinematics are well established, and have been applied in a broad range of cognitive and visuomotor tasks. Yet specific knowledge regarding FPS performance is lacking. Visuomotor skills are specialized and are typically constrained to the contexts and modalities in which they are learned. Thus, previous visuomotor psychophysics research provides limited knowledge about FPS performance.
A system and method may measure or calculate performance or motor acuity, a visuomotor skill performance metric, in a person playing a computer task, gamified task, or computer game by determining, for a number of different target types, speed of the player and one of precision, accuracy and variability of the player (e.g. in movement or shots); and determining or calculating motor acuity based on the relationship between speed of the player and the one of precision, accuracy and variability of the player. The calculation may be using a statistical analysis; and may be by creating a best-fit line in a comparison of speed and variability; and transforming the best-fit line to a curve in a comparison of player speed and player precision (1/variability).
Some embodiments evaluate the impact of hardware, software, and various other factors (e.g., mouse, mouse pad, rendering latency, posture, exercise, diet and dietary supplements, sleep, etc.) on gaming performance.
Some embodiments are applicable to other fields. Examples include: rehabilitation (assessment and training) for patients recovering from stroke, traumatic brain injury, or arm injury; concussion assessment, to determine if an individual may have suffered a concussion or brain injury, by comparing their performance with a baseline that was measured before the incident that may have caused an injury, or to determine if an athlete should return to play following a concussion; assessment and training for individuals who are learning to use a prosthetic limb; assessment and training for patients with movement disorders (e.g., cerebral palsy); gamified assessment of sensorimotor and cognitive fitness, e.g., immersive games for military personnel that may run on mobile devices (phone, tablet, etc.) and that assess and monitor changes in cognitive fitness, readiness, and performance across the military lifecycle and in the field.
The subject matter regarded as the invention is particularly pointed out and distinctly claimed in the concluding portion of the specification. The invention, however, both as to organization and method of operation, together with objects, features, and advantages thereof, may best be understood by reference to the following detailed description when read with the accompanying drawings in which:
It will be appreciated that for simplicity and clarity of illustration, elements shown in the figures have not necessarily been drawn to scale. For example, the dimensions of some of the elements may be exaggerated relative to other elements for clarity. Further, where considered appropriate, reference numerals may be repeated among the figures to indicate corresponding or analogous elements.
One skilled in the art will realize the invention may be embodied in specific forms other than the examples presented herein without departing from the spirit or essential characteristics thereof. The foregoing embodiments are therefore to be considered in all respects illustrative rather than limiting of the invention described herein. Scope of the invention is thus indicated by the appended claims, rather than by the foregoing description.
In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those skilled in the art that the present invention may be practiced without these specific details. In other instances, well-known methods, procedures, and components have not been described in detail so as not to obscure the present invention. Some features or elements described with respect to one embodiment may be combined with features or elements described with respect to other embodiments. For the sake of clarity, discussion of same or similar features or elements may not be repeated.
Although embodiments of the invention are not limited in this regard, discussions utilizing terms such as, for example, “processing,” “computing,” “calculating,” “determining,” “establishing”, “analyzing”, “checking”, or the like, may refer to operation(s) and/or process(es) of a computer or other electronic computing device, that manipulates and/or transforms data represented as physical (e.g., electronic) quantities within the computer's registers or memories into other data similarly represented as physical quantities within the computer's registers and/or memories or other non-transitory storage medium that may store instructions to perform operations and/or processes.
Embodiments may assess video game inputs to measure performance metrics, e.g. flicking skill or motor acuity for a person for example based on flicking movements, and movement kinematics performance metrics of esports players. Player activity as described herein (e.g. flicking, firing a shot) is typically provided as input to a computer system, the player typically viewing a computer-generated display. A player “moving” may refer, for example, to a user providing input, e.g. using a game controller or computer mouse (e.g., input device 135 or 205) which causes some displayed representation of that user to move, or to move the field of view.
In some embodiments, the orientation of the player's view of the environment, controlled by the input controller or device, is recorded in Euler angles and sampled, e.g. at 120 Hz. Other sampling rates can be used instead of 120 Hz; one embodiment is implemented with sampling rates ranging from 30-240 Hz. Other representations of player orientation and movement can be used instead of Euler angles, such as pixels or quaternions. A game entity such as an aiming image or icon, e.g. a crosshair, may be marked in some embodiments with a dot at the center of the screen, and may correspond to the direction in which a shot would be fired, and may be used to determine precision, accuracy or error. When triggered (e.g. when the user commands the system to fire a shot using an input device or controller), the shot may move, in the virtual environment of the game, in a straight line pointing away from the player's virtual avatar. The shot may then strike an entity such as a target, and thus the relative game or screen position of the aiming image and target may be used to determine precision, error or accuracy.
Some embodiments utilize or take input from products from Aim Lab, such as Aim Lab's first-person shooter (FPS) assessment and training game, to measure using one or more games or gamified tasks that involve shooting targets with a number of different of target types (e.g. target sizes) and/or target presentation durations. Human performance is known to exhibit a speed-accuracy tradeoff. The different target sizes and durations incentivize players to make faster and less precise shots, or to make slower and more precise shots. The adaptive features of some gamified tasks, e.g., titrating target presentation durations and/or target sizes, may result in an assessment with no “ceiling effect”.
Embodiments of the present invention may provide improvements over prior technology used for assessing motor acuity, or game performance. Prior art methods of assessing player performance may not be as effective as those of embodiments of the present invention. One prior method of characterizing visuomotor performance is the use of Fitts Law for a wide range of tasks including an FPS task. Fitts Law predicts that the time required to rapidly move to a target is a function of the ratio between the distance to the target and either the size of the target or the spatial error of the movement. Motor acuity, complementary to Fitts Law, may be defined as the ability to execute actions more precisely and within a shorter amount of time. In one example Fitt's Law application to FPS gaming was based on measuring shot time (e.g., elapsed time between the destruction of a target and the first shot at the subsequent target), target distance (e.g., angular distance to the target when it first appeared) and shot spatial error (e.g., angular distance between the shot and the center of the target) for two gamified tasks, each task including one of the two target sizes. Performance, combined across the two target sizes, was poorly characterized by Fitts Law. In one example, for each target the shot time, target distance and shot error were measured. The logarithm of the ratio of target distance and shot error was computed, separately for each target, and then binned by decile. For each bin was computed: 1) the median shot time; and 2) the median logarithm of the ratio of target distance and shot error. Fitts Law provided a good characterization of behavioral performance when fit separately to the data from each task (e.g., separately for each target size), but a very poor characterization of the data when combined across the two tasks (e.g., combined across the two target sizes). In other words, shot times were well predicted by Fitts Law as a function of target distance and target error for each target size separately, but not for both target sizes together.
This failure of Fitts Law may reflect a difference in strategy between the two tasks, e.g., that players tended to be more “swipy” for large targets than for small targets. A previous study found excellent agreement between Fitts Law and behavior in an FPS task, but the participants in that study were either amateur gamers or did not report playing FPS games at all. Swiping is an advanced skill that is commonly used by highly competitive and professional players when speed is of paramount importance.
Embodiments of the present invention may provide a better method to measure skilled performance, one that isolates skill from the player's strategy. Some embodiments may assess skill or motor acuity as a relationship of a speed metric and a variability or precision metric. Each of speed and variability or precision may be measured based on player input such as movement kinematics performance metrics (e.g. median movement speed, median movement landing error) or shot performance metrics (e.g. median elapsed time to first shot and median first shot error), typically multiple data points being used to calculate speed and variability (and typically multiple data points of one target type being used to compute speed, precision, variability and/or accuracy). Precision may be 1 divided by variability.
Embodiments may measure performance metrics, such as motor acuity in a game, based on a speed/precision or speed/accuracy tradeoff. In some embodiments this is based on user input for a plurality of trials, with a plurality of different target types or sizes, where some trials may use the same target type, e.g. 100 trials using two target sizes, where 50 trials use target size A and 50 use target size B. Some embodiments rely on evaluating a player on two or more different tasks, typically performed back-to-back, or close together in time (e.g. on the same day, in the same hour, etc., as the player's skill may change over longer time periods), each task typically defined by a different target type or size. Each task may have distinct incentives regarding speed and precision. Embodiments may evaluate a player in isolation (e.g. not against other players) or against other players. Embodiments may evaluate a player by having the player perform a game or gamified task using a first target type (e.g. large target) then perform a game or gamified task using a second target type (e.g. smaller target), and possibly repeating with other target types. Some embodiments may interleave (e.g., in randomly shuffled order) the different target types or sizes in a single game or gamified task. Some embodiments may measure performance metrics, e.g., motor acuity, by analyzing or plotting the relationship or tradeoff between speed and accuracy, variability or precision in the tasks, e.g. by measuring the distance of a curve or function plotting speed vs. precision from the origin of a graph. Such a curve or function may represent a speed-accuracy tradeoff (SAT). Some embodiments perform a statistical analysis to determine if there is a statistically significant change or difference in speed and accuracy, variability or precision, e.g., between players, over time (e.g., with practice or training) or across different conditions (e.g., impact of hardware, software, and various other factors such as mouse, mouse pad, rendering latency, posture, exercise, diet and dietary supplements, sleep, etc.) on gaming performance.
Embodiments may determine motor acuity or other performance metrics based on a tradeoff or comparison between performance metrics such as movement speed and movement variability, movement precision or movement accuracy. In one embodiment, this may be done for a number of different target types or sizes: the player may perform a certain task, e.g. the same task, more than once, each time for a different target type or the different target types may be interleaved during each round of the task. For data acquired with two or more target types a statistical analysis may be performed or a curve generated, analyzing speed/accuracy relationship or tradeoff, speed/precision relationship or tradeoff, or another relationship. (Some tasks may include only one target type; other tasks may interleave different target types in the same task.) A curve, e.g. an SAT (speed-accuracy tradeoff) curve, may be generated: the curve may represent the possible combinations of movement speed and movement precision given the player's skill (typically measured on a certain day or certain session, local in time), with the player's current strategy determining where along the curve their current speed and precision is drawn from. Embodiments may measure both speed and another performance metric such as accuracy, variability or precision; use of a single performance metric as a proxy for motor acuity may be less effective, as motor acuity may be defined by both speed and accuracy. Further, a line fitted to a player's movement speed and movement variability may be used to help a player adjust strategy: if the slope is negative, near zero, or near vertical, this indicates that the player did not make the trade-off between speed and precision to maximize performance between scenarios (e.g., target types). In some embodiments such an analysis may be done for a number of players, to rate or rank the players against each other. Embodiments may use such statistical analysis or graphing to compare the performance of a player (a “newly seen” player) to that of other players stored in a database; to that player in a different day (e.g. to gauge improvement or decline, or to gauge performance of a player over time).
Embodiments may likewise determine motor acuity based on shot performance metrics such as shot speed (e.g., 1 over median elapsed time between target appearance and the first shot toward each target) and shot variability (e.g., variability of shot locations relative to the centers of the targets), shot precision or shot accuracy.
In one example three different target sizes may be used: e.g. small target size; medium target size; and large target size. Shot speed (e.g., 1 over median elapsed time between target appearance and the first shot toward each target) and shot variability (e.g., median angular shot error wherein shot error is the absolute value of the angle between the shot and the center of the target) may be measured, separately for each target size and individually for each player. For each player a line may be determined as a best fit line to the shot speed (e.g., in units of seconds−1) versus shot variability (e.g., in units of degrees) data from one player. Different best fit lines may be determined for each player. Such lines may be re-plotted as curves in terms of shot speed versus shot precision (in this example 1 over shot variability in units of degrees−1). The resulting curves represent the transformed lines from each player. A diagonal dashed line may extend from the origin (0,0) to a multiple of (e.g., three times) the standard deviation plus the mean of the distribution of shot precision (x-axis) and shot speed (y-axis) across curves from all players. In some embodiments, the intersection of individual curves with the diagonal, indicates flicking skill: the further the intersection from the origin, the higher the skill.
Players may make multiple movements and/or multiple shots to destroy a target. Speed may be measured as the median movement speed for each of the first movements toward each target (e.g., as opposed to the use of movements toward each target subsequent to the first movement). Variability may be measured as the variability of the movement landing points for the first movements toward each target (e.g., as opposed to the use of movements toward each target subsequent to the first movement). Speed may be measured as the shot speed (e.g., 1 over median elapsed time between target appearance and shot fired) for the first shot toward each target (e.g., as opposed to the use of shots subsequent to the first shot). Variability may be measured as the variability of the first shot to each target (e.g. as opposed to the use of shots subsequent to the first shot).
In some embodiments of the present invention, flicking skill assessment (FSA) may be calculated as a performance metric, which may be a robust and sensitive measure to assess and compare skill between players and to track improvement over time of a player's skill as they continue to train and practice.
Some embodiments characterize fine-grained kinematics of players' movements: speed, precision, accuracy, reaction time (SPAR), and swipiness (e.g., tendency to shoot while in motion rather than landing on a target before shooting).
Some embodiments for measuring motor acuity and movement kinematics performance metrics were validated by collecting and analyzing performance data from professional and amateur esports players. The results of this analysis revealed that: 1) movement kinematics depended on task demands (e.g., reaction time was shorter, and movements were less precise and more “swipy”, for tasks that incentivized speed over precision); and 2) individual differences in motor acuity were highly correlated with both movement efficiency (number of movements needed to hit a target) and movement kinematics (reaction time, precision, and swipiness).
Embodiments of the present invention include methods to measure and characterize ballistic or other movements, for example during FPS game scenarios. Ballistic movements may be planned then executed without making any adjustment during the movement, and without using sensory feedback, so the movement is as fast as possible; this is in contrast to some human movements that are made more slowly, using sensory (e.g., visual or tactile) feedback to fine tune the movement while making the movement.
Human performance, e.g. motor acuity, exhibits a speed-accuracy tradeoff: the speed at which a response or action is made is negatively correlated with the accuracy or precision of that action. People can be very fast and less accurate, very accurate and slow, or somewhere in between. This effect can be evident for different aspects of speed (e.g., reaction time, movement speed) and different aspects of accuracy (e.g., percent of correct responses/decisions, movement accuracy and variability). A hallmark of speed-accuracy tradeoff or relationship is the ability of humans to adapt to current demands and prioritize speed and accuracy relative to each other. If a task requires very fast responses or movements, a person may sacrifice accuracy to maximize speed. If there is a high cost to incorrect responses, a person may take longer to respond or move more slowly to maximize accuracy.
In FPS performance, speed-accuracy tradeoff or relationship can be characterized in terms of the movement kinematics performance metrics (e.g., reaction time to initiate a movement, movement speed, movement accuracy, variability or precision of a plurality of movements). Speed-accuracy tradeoff can be characterized also in terms of shot performance metrics (e.g., the time interval between when a target appears and when the first shot is fired and the spatial error of shots). Embodiments may use performance metrics computed based on shot performance. Some embodiments may include or evaluate movement kinematics performance metrics. Differences in flicking skill—and resulting placement of speed-accuracy tradeoff curves—arise between players of different individual skill, as well as within players as they learn and improve or perform under different conditions.
Performance on a single task with a single incentive for speed vs. accuracy may be insufficient to estimate a player's motor acuity or FSA metric, as the speed-accuracy tradeoff cannot be characterized from a single data point. It may be necessary to assess performance in a number of different conditions with different priorities of speed and accuracy. Embodiments may isolate an individual player's skill from their chosen trade-off between speed and accuracy, and hence may represent a sensitive measure of skill differences and improvements rather than bias or strategy.
Some embodiments use products or systems such as Aim Lab systems, an FPS video game. Aim Lab's systems include a large number of tasks that replicate gaming scenarios.
In addition to typical flick-and-land movements, players sometimes choose to maximize speed by shooting “on the fly” instead of slowing down before firing, a tactic referred to as a “swipe” movement. For swipes, the endpoint of the movement may be well past the target location. Some embodiments characterize whether movements more closely resemble a typical swipe versus a typical flick-and-land movement. There is a continuum of “swipiness” with the ideal flick-and-land and the ideal swipe corresponding to two ends of the continuum. In the process of validating an embodiment using an FSA metric, it was found that movement kinematics depend on task demands: reaction time is shorter, and movements are less precise and more “swipy”, for a task that incentivizes speed over precision.
In one embodiment an FSA metric may improve evaluation technology by providing highly sensitive and robust estimates of motor acuity performance. Motor acuity is the ability to execute actions more precisely, and/or within a shorter amount of time. There is a paucity of studies examining motor acuity, a gap likely linked to the tight resource constraints on laboratory-based studies. The handful of lab studies that examine motor acuity have used relatively simple motor tasks, like drawing circles as fast as possible within a predefined boundary, throwing darts, or center-out reaching and grasping.
Embodiments of the invention may evaluate performance based on data produced by a player playing games or gamified tasks such as Aim Lab products (e.g. Sixshot, Gridshot or Adaptive Reflexshot tasks); or other tasks such as tasks specifically designed for therapy. In order to develop one embodiment, a large sample of Aim Lab data (over 7,000 players and over 60,000 repeats of the 60-second Gridshot task) over a period of months was used to examine motor learning. In this database, hits per second was used as a proxy for motor acuity. Embodiments of the present invention may instead use an FSA metric for measuring motor acuity in FPS games.
Aim Lab software systems include commercial software products written in the C# programming language using the Unity game engine. Unity is a cross-platform video game engine used for developing digital games for computers, mobile devices, and gaming consoles. Players may download an Aim Lab system, e.g. to their desktop, laptop or PC, etc. While Aim Lab systems are one example of systems that may be used with embodiments of the present invention, other systems may be used.
Players may control their virtual weapon or other virtual tool in Aim Lab tasks using input devices or controllers such as a mouse and keyboard while viewing the game on a computer screen. Performance data or measurements of user's performance may be uploaded to Aim Lab secure servers. Aim Lab systems includes a large number of different task scenarios for skill assessment and training, each tailored to a facet of FPS play. These task scenarios may assess and train users in the areas of, for example, visual detection, motor control, tracking moving targets, auditory spatial localization, change detection, working memory capacity, cognitive control, divided attention, and decision making. Each task can be customized to prioritize accuracy, speed, and other basic components of performance. During a round, players may be granted points for each target that they successfully shoot and destroy. In some tasks, additional points may be rewarded for targets destroyed more quickly. Players attempt to maximize their score on each round by destroying as many targets as possible. Different task scenarios may incentivize players to prioritize accuracy over speed or to prioritize speed over accuracy.
Player input may include manipulation of the input device (e.g. mouse, Wii controller) to move a player, or move a player's aim (or the aim of a weapon), or the view of the player; and also trigger or shot input to fire a device; and also input to move the player in a game. For example a user may move a mouse to alter the aim of the player, and click on a mouse button to fire a weapon in a game. Measures of speed, precision, accuracy, swipiness, etc. may be based on the relationship between game entities (e.g. targets, player point of views (POVs), the aim of a weapon), the movement of these entities (e.g. the aim of a weapon moving, a target moving), and the timing of entity actions (e.g. a weapon firing, a target appearing, etc.). For example, the accuracy of a shot triggered by user input to a mouse or game controller may be determined by the point of aim (e.g. crosshairs) at the time of the shot, relative to the point or position of the intended target. A user may move their avatar or player by, for example keyboard input (e.g. the known arrow key or WASD input method), which in turn may affect the game entity of the player's point of view, which is often in FPS games coextensive with the entity of the crosshairs or aim of the virtual weapon. The player may use the input device to aim the weapon, which typically means aiming or rotating the POV, and may provide trigger input to the input device to fire the weapon. Other methods of user input may be used. The trajectory or movement of the aiming or crosshairs may be used to measure player movement kinematics performance metrics (e.g. movement speed, movement precision, movement accuracy, movement reaction time, or swipiness). The crosshairs or aiming position of the weapon may determine where the shot lands: for example if the crosshairs are pointed at/displayed over the center of the target when the shot is fired, the shot is determined to be or deemed to be at the center of the target.
Some embodiments of the present invention utilize task scenarios that assess flicking skill. Flicking may be described as sudden and rapid ballistic movements by a player in a game (both the sudden and rapid motion of the input device and the on-screen movement), for example to aim and fire a weapon to destroy (e.g. stationary or moving) targets. For example, embodiments utilize flicking tasks with different target types (e.g. sizes and/or different target presentation durations), to characterize each player's speed-accuracy trade-off and to estimate their flicking skill.
In some embodiments, human participants or players play FPS task(s) remotely using their own gaming set-up. A gaming set-up may include for example hardware such as a PC, or other computer, monitor, mouse and mouse pad, and may operate based on settings (e.g., stored at user computer 200,
In some embodiments, movement trajectories (e.g., of game entities, objects or on-screen objects such as an aiming point, a weapon, etc.) may be determined by sampling an input position from an input device or controller, controlled by a person while the person plays the game. The input controller can control a number of features of the player, including for example movement of the player, aiming of a player's weapon (e.g. signaled by crosshairs) and/or one or more actions of a player (e.g., utilization of a weapon, such as firing a shot). Examples of input controllers used with some embodiments include keyboard, mouse, gaming mouse, video game console, joystick, accelerometer, gyroscope, pointing device, motion capture, Wii remote controller, eye tracker, computer vision system, or any of a variety of methods, devices, apparatuses, and systems for sensing, measuring or estimating human movement to provide input signals to a computer. For example, the gyroscope in a smart phone or tablet can be used as the input device for games or gamified tasks that run on those devices. Examples of human movements used with some embodiments include hand movements, arm movements, head movements, body movements, and eye movements. Examples of input controllers also include brain-computer interface methods, devices, apparatuses, and systems for sensing, measuring or estimating brain activity. Examples of brain-computer interfaces include, but are not limited to, devices that measure electrophysiogical signals (e.g., using electroencephalography (EEG), magnetoencephalography (MEG), microelectrodes) and optical signals (e.g., using voltage-sensitive dyes, calcium indicators, intrinsic signals, functional near-infrared spectroscopy, etc.). Examples of brain-computer interfaces also include other neuroimaging techniques (e.g., functional magnetic resonance imaging). Examples of input controllers also include methods, devices, apparatuses, or systems for sensing, measuring or estimating physiological data. Physiological data includes, but is not limited to, EEG, electrocardiogram (EKG), electromyography (EMG), electroocoulogram (EOG), pupil size, and biomechanical data relating to breathing and/or respiration. A person of skill in the art recognizes that any such input controller or any combination of such input controllers could be used. It is also recognized that other methods, devices, apparatuses, or systems for sensing, measuring or estimating human movement or physiological activity could be substituted for use with some embodiments, including those that have not yet been reduced to practice.
Some embodiments evaluate a user or player (e.g. evaluate a skill such as performance in flicking skill tasks) by having the player participate in or provide input to one or more of games or gamified tasks or computer games, such as a Gridshot task, a Sixshot task, an Adaptive Reflexshot task or other tasks. Data gathered from these tasks may be used to determine performance metrics in the relevant computer game tasks, for example by presenting, in the relevant gamified task, targets of varying types (e.g. varying sizes), gathering data such as variability, precision, accuracy, speed (e.g. time from target presentation to first shot). A person of skill in the art recognizes that any of a wide variety of other tasks can be substituted.
For example,
Performance metrics or data may be gathered from a player's use of a Sixshot task (an example of which is shown in
Performance metrics or data may be gathered from a player's use of a task which presents the user with a plurality of target types (e.g. sizes, and/or target presentation durations, or other types), with the target size and/or duration adjusted adaptively to match the player's performance at any given time point. Adaptive Reflexshot (an example of which is shown in
Some embodiments may evaluate performance based on movement kinematics performance metrics such as movement amplitude, movement speed, accuracy of movement landing point, variability or precision of movement landing point, reaction time between target appearance and initiation of the movement and swipiness. Various measures of speed, accuracy, variability or precision of movements, and other values based on movement kinematics, may be used in different embodiments. Some embodiments may analyze movement kinematics, and for example fit parametric functions to the time series of each player's movements.
Some embodiments measure movement kinematics of each player for one or more tasks. Players often aim for a target by initiating a movement toward the target, increasing and then decreasing movement speed, and firing a shot after slowing down or coming to a full stop. In some embodiments, the kinematics of these movements are characterized by fitting a sigmoid to the movement time-series. The best-fit parameter values indicate or are a measure of movement accuracy (e.g. amplitude of the sigmoid), movement speed (e.g., slope of the sigmoid), movement precision (e.g., 1 over (divided by) the median absolute difference between the sigmoid endpoint of the movement and the center of the target, for a plurality of movements), and reaction time (e.g., time point at which the sigmoidal movement was initiated).
For example, player orientation (e.g. orientation of crosshairs or weapon) during each period labeled as “in-motion” (e.g., by movement parsing) may be fit with a sigmoidal function, for example:
where example Equation 1 defines a sigmoidal function. In this function a, b, and c are the parameters of the sigmoid that determine the lift off (which indicates reaction time), the slope (which indicates speed), and the amplitude (which indicates accuracy). The values of x(t) in example Equation 2 represent a model of the horizontal component (rotation about the y-axis) of the movement trajectory for each time sample, where p1 becomes a, p3 becomes b, and p4 becomes c. The values of y(t) in example Equation 3 represent a model of the vertical component (rotation about the x-axis) of the movement trajectory. Example illustrations of the model components x(t) and y(t) are shown in
Accuracy may be how close a given set of measurements (observations or readings) are to their true value, while precision may be how close the measurements are to each other. Precision may be a measure of the consistency with which a player performs: e.g. a player consistently landing a shot a certain distance to a target indicates precision, while a player sometimes landing a shot close to and sometimes far from a target indicates less precision. Accuracy, on the other hand, may measure how close to a target a player's shots hit (e.g., on average): systemically being close to a target indicates high accuracy and systemically landing shots far from a target indicates lower accuracy.
In some embodiments, speed, accuracy, and reaction time may be calculated from, or measured based on, the best-fit parameter values, for example:
1. Speed (e.g., cm/sec or deg/sec): peak speed at the midpoint of the movement, e.g. the movement of the user-controlled game entity such as the weapon, aiming point, or crosshairs.
2. Accuracy (e.g., % distance of the landing point of the movement to the target): which may be described as distance (e.g. spatial error) between the landing location of the movement and the center of the target. For example if a 20 degree movement is required but the player moves 18 degrees then the accuracy is −10% (10% short of the target) whereas if the player moves 22 degrees then the accuracy is +10% (10% past the target). Distance may be measured in e.g. an angle, pixels, a virtual distance having meaning in the reality displayed in the game, actual distance of an input controller to the distance of that controller corresponding to a target (e.g., mouse position on the mouse pad), or other measures.
3. Reaction time (e.g., sec): time interval between when the target appeared and the initiation of the movement (e.g., when the movement reached 5% of its endpoint).
Movement speed and accuracy of each parsed movement can be quantified as for example:
The position of the player (e.g. the position of the crosshairs) at time t is represented as xt and yt. The function f′(t) is the derivative of the sigmoidal function (Equation 6). The value of am represents the amplitude of the movement (Equation 7) and the value of at represents the distance to the target location (Equation 8). The vector e represents the movement error (Equation 9) and the vector u represents a unit vector in the direction of the target location (Equation 10). In some embodiments, accuracy is calibrated to have units of cm and/or movement speed is calibrated to have units of cm/sec. In some embodiments, accuracy is calibrated to have units of deg (e.g., angular degrees of rotation in the virtual environment of the game) and/or movement speed is calibrated to have units of deg/sec.
Other units, and other ways of calculating speed, accuracy, and reaction time, may be used. Speed, accuracy, and reaction time may be measured based on multiple movements, e.g., by computing the average or median across a plurality of movements.
Movement precision may be computed as 1 divided by movement variability. Movement variability may be computed as the standard deviation (across a plurality of movements) of movement landing points. Movement variability may instead be computed as the median (across a plurality of movements) of the absolute value of the spatial error (e.g., distance between the movement landing point and the center of the target). Variability may be expressed in units of % distance and precision may be expressed in units of 1/% distance. Other measures of variability may be used instead of standard deviation and median absolute error and variability may be expressed in units other than percentages (e.g., angular degrees, centimeters).
To characterize “swipiness”, the degree to which each ballistic movement resembles a swipe (e.g. shooting on the fly) versus a flick-and-land (e.g. slowing down and stopping before firing) movement, some embodiments compare the time of each shot with the time of the midpoint of the corresponding ballistic movement. An ideal swipe corresponds to firing a shot at the midpoint of a movement, e.g., the time point with maximum speed. An ideal flick-and-land corresponds to firing a shot only after the movement has ended, after the midpoint of the movement. Some embodiments compute swipiness as a ratio of the time of the shot and the time of the midpoint of the movement (p4), divided by 2. This results in a “swipiness” value of 0.5 (example arbitrary units) for an ideal swipe, and a “swipiness” value more than or equal to 1 for an ideal flick-and-land. Other thresholds, and other formulas, may be used. In these embodiments, there is no swipiness value computed for a movement with no associated shot, e.g., if there is no shot between the initiation of one movement and the initiation of the next movement. Swipiness is a granular measure of shot speed as it relates to the movement trajectory, with lower swipiness indicating that the firing of a shot occurred earlier in the trajectory. Additionally, the lack of a swipiness value indicates that a movement is not associated with a shot. Moreover, across trials within a certain context or task, the number of movements with no swipiness value reflects the need for players to make multiple movements to destroy any given target.
Embodiments may measure speed, accuracy, or precision of shot performance; other values based on shot performance may be used in different embodiments. Some embodiments operationalize or calculate speed and precision according to the first shot fired at each target. For example, speed can be computed by or be a measure based on: 1) determining the elapsed time between target spawn or appearance and the first shot fired (e.g. an input to fire a shot received from a game controller), separately for each target; 2) computing the mean or median elapsed time for this spawn-shot time across a plurality of targets; and 3) computing 1 divided by that mean or median elapsed time. Likewise, precision can be computed by: 1) computing the shot error of the first shot that was fired at each target (e.g., the distance between the landing point of the shot in the game and the center of the target); 2) computing the mean or median of the absolute value of the shot error across a plurality of targets; and 3) computing 1 divided by that mean or median shot error. Other methods of calculating shot speed, precision or variability, may be used for a plurality of shots.
Distance may be measured at the time a shot is taken. A shot being taken may include that a player provides input to an input device such as a mouse or game controller indicating that the player wants to fire a virtual weapon (e.g. rifle) at a target displayed in a game. Distance may be measured between the landing point of the virtual shot in the game and the center of the target.
Shot performance or movement kinematics may be used to determine precision, variability, or other measures. In some embodiments multiple measures of shots or movement kinematics (typically for a specific target type)—typically error—may be used to calculate precision or variability. For example, multiple measures of error for a metric (e.g. shot error or movement error) may be taken; variability may be computed for this error; and precision may be determined as 1/variability (thus variability may be 1/precision). Error in some embodiments may be absolute value of error, e.g. absolute value of distance between target center and movement or shot landing point, or other error. Precision (e.g. shot precision, movement precision) may be calculated as or be a measure based on, in one embodiment 1 over (divided by) the variability. A person of skill in the art recognizes that there are a variety of alternative ways to compute shot or movement variability. Some embodiments compute shot variability as the standard deviation of shot locations or movement landings, or standard deviation of accuracy over a plurality of movements or shots. Some embodiments instead compute the median absolute deviation (MAD) of measures such as movement or shot accuracy, or the median of the absolute difference from the median. MAD is a robust measure of variability and may minimize the impact of outliers.
Precision or other measurements may be based on one player's multiple shots, movements or trials, or multiple players multiple shots, movements or trials, e.g. in order to compare one newly seen player to multiple other players in a database, or one player against that same player's prior performance. A database may be created of speed, precision or variability, accuracy or error, reaction time and other performance metrics for players (e.g. past players) for comparison to new or future players.
A variety of measures of variability could be used instead of standard deviation or MAD. Some embodiments compute the square-root of the median squared difference from the median of, e.g. shot accuracy or movement landing accuracy, or of squared movement error, square root of the median of the squared deviation, or other player measures. Embodiments compute shot variability based on shot error, e.g., the distance between each shot location and the target center (rather than the mean or median of the shot locations). Embodiments can compute the mean of the movement or shot errors across a plurality of shots or movements, median of the errors, the square-root of the mean, or of the median, of the squared errors, or the square-root of the median of the squared errors. Embodiments may compute movement variability based on movement error, e.g., the distance between the landing point of each movement and the target center (rather than the mean or median of the movement landing points). Embodiments can compute the mean of the movement errors across a plurality of movements, median of the movement errors, the square-root of the mean of the squared movement errors, the square-root of the median of the squared movement errors.
For task scenarios that titrate target size, some embodiments operationalize variability and precision based on a target size threshold. A target size threshold may be computed by for example by: 1) determining for each target the target size (e.g. in pixels on a display; degrees of visual angle; or another measure) and whether or not the target was hit; 2) calculating the hit rate (proportion of targets hit) for each of the possible target sizes; and 3) fitting a parametric function to the resulting hit rates. Some embodiments use a maximum likelihood fitting procedure. Other methods for fitting psychometric data are well-known to a person of skill in the art. A target size threshold may instead be determined as the target size value converged on during titration. Precision can then be computed as 1 divided by the size threshold.
Embodiments may use a variety of ways to compute speed of shots or movement kinematics. Embodiments can compute, or measure speed based on, for example, shot speed (e.g. shots per second), hit rate (e.g., in units of hits per second), movements per time period (e.g. second), mean or median movement speed (e.g., in units of cm per second, degrees per second, pixels per second, etc.), peak movement speed for a parsed movement, the slope of a sigmoid describing a movement, mean or median movement reaction time, mean or median shot time (time interval between target appearance and first shot), or other measures.
For task scenarios that titrate target duration, some embodiments operationalize speed based on a target presentation duration threshold. A target presentation duration threshold may be computed by for example: 1) determining for each target the presentation duration and whether or not the target was hit; 2) calculating the hit rate (e.g., proportion of targets hit) for each of the possible presentation durations; and 3) fitting a parametric function to the resulting hit rates. Some embodiments use a maximum likelihood fitting procedure. Other methods for fitting psychometric data are well-known to a person of skill in the art. A target presentation duration threshold may instead be determined as the target duration value converged on during titration. Speed can then be computed as 1 divided by the duration threshold.
According to some embodiments, speed is operationalized as 1 divided by target presentation duration threshold (speed may be a unitless measure in such embodiments). Shot speed (in example units of 1/seconds) may be compared to a function of the shot error (in example units of degrees) for the Adaptive Reflexshot task with multiple (e.g. three) different target types or sizes. Shot speed may be for example computed from the target presentation duration threshold; and shot error may be for example computed from the distance between the center of a target and the first shot that was fired at that target.
A parsing or other process may segment movement data into individual movements so as to measure movement kinematics such as speed, reaction time, and movement accuracy separately for each individual movement. An individual movement may be defined from the stream of input from a user, for example by parsing.
A primary movement of an aiming point, a player, or a player's weapon toward a target is frequently followed by a corrective movement. For instance, the primary movement can be hypermetric, passing beyond the target, with a corrective movement in the opposite direction back toward the target (see example in
Some embodiments parse the time-series of each player's movements acquired by the input device or controller. This may take as inputs a number of player inputs over time and produce, as an output, a number of distinctly defined movements. Movements can be recorded as change in orientation (e.g., Euler angles) in the virtual environment of the game. Movement parsing may label each time-point as for example in-motion or stationary. Epochs (e.g., periods of time corresponding to a consecutive series of time samples) are labeled as in-motion after one or more consecutive samples exceed a velocity threshold. Epochs are labeled as stationary after one or more consecutive samples fall below a velocity threshold. The number of consecutive samples used and the values of the thresholds can be determined from a large dataset comprising data from a large number of players. Each successive epoch (e.g., in-motion or stationary) may be different in duration.
Players may change the direction of their movements without significantly slowing their movement speed. For example, this may occur when multiple targets are present in the virtual environment, and a player begins to move towards one and then decides to prioritize a different target without destroying or shooting at the original target. As another example, a player may not reach a target quickly enough, and thus will never fire a shot at that target even though a ballistic movement was initiated in that target's direction. Some embodiments utilize changes in movement direction to parse movements. If the player moves in one direction and then changes to a substantially different direction, then the time-series may be segmented such that an epoch preceding the change in direction are associated with one movement and the epoch following the change in direction are associated with a different movement. Some embodiments detect a change in direction when the angle of the movement velocity changes above a threshold for one or more consecutive time samples. The number of consecutive samples and the values of the thresholds can be determined from a large dataset comprising data from a large number of players. An example is shown in
In some embodiments, performance metrics can include the median number corrective movements to hit a target and the proportion of targets for which the player hits the target with N movements (where N is a positive integer).
Some embodiments associate each of a player's movements with a shot. Players typically make a sequence of movements to destroy a target. To associate a shot with one or more movements, some embodiments first identify the time point at which the shot was fired, and step backwards in time (e.g., asses) from the time at which a shot was fired to identify the movements that were initiated at time points preceding that shot. Each shot may be associated with more than one movement. Each movement may be associated with more than one shot. For example, a player may fire a shot that misses the target during an initial swipe movement, followed by a corrective movement and a shot that destroys the target. In this example, the initial swipe movement may be associated with both shots. Some embodiments associate movements with corresponding targets according to which shots are associated with that movement and which target or targets was or were closest to the associated shot or shots.
Performance metrics in some embodiments can include the proportion of movements for which there is no shot associated with the movement, which reflects the need for players to make multiple movements to destroy any given target.
Data analysis and validation may be performed in varying manners. For example, input controller or device calibration may be performed.
Some embodiments measure the player's physical movements (e.g. the movements of a mouse on a mousepad when the input controller is a computer mouse). To do so, the orientation data of the player as depicted in the game (e.g. the crosshairs) may be converted to the corresponding movements of the mouse on the mouse pad (e.g. the physical movements that were used to generate player movement), e.g. in units of centimeters or inches. This may require additional information about the relationship between physical mouse movement and changes in player orientation, which can vary across players due to their hardware and software settings. Other input devices may be used.
Some embodiments record the in-game settings that govern the field of view as well as the mouse or input device sensitivity, e.g. the magnitude of a change in camera rotation of field of view displayed to a user (which may determine aiming, crosshairs, or a weapon position) that results from a single increment or count from the mouse. Mouse sensitivity can be recorded for the x- and y-axes of motion independently. Additionally, some embodiments record a player's mouse (e.g. hardware, software, or both) settings that determine the number of counts that result for one unit (e.g. inch) of distance traveled (e.g., counts per inch, CPI).
In one embodiment, player camera orientation or aim point (e.g. position of cross hairs over a target) is converted into units (e.g. centimeters) of physical mouse movement, for example according to a formula such as:
1. Mouse sensitivity*0.05=angle increment (degrees turned per count)
2. Total degrees turned/angle increment=counts
3. (counts/CPI)*2.54=physical distance traveled (cm)
The value of 0.05 is an example constant used in the Unity software code to scale mouse counts to degree increments; other constants, other formulas, and other software tools may be used. In other embodiments the input controllers can be calibrated using similar methods and computations, that distances can be specified in metric units (e.g., centimeters) imperial units (e.g., inches) or other units.
Data post-processing may be performed, e.g. after obtaining data from games or gamified tasks or computer games and, e.g., prior to using the data to evaluate motor acuity or flicking skill. Some embodiments apply one or more steps of post-processing to movement kinematic metrics (e.g. speed, precision, accuracy, reaction time, and swipiness) or shot performance metrics to remove outliers and/or to decorrelate the values. Some embodiments apply an upper and lower band threshold to remove outliers (e.g. presumed to be failures in the trial parsing or sigmoid fits) from the movement kinematic metrics. For example, the lower band threshold can be 0 and the upper band threshold can be and the 95th percentile. In some embodiments, movements are pruned (e.g. ignored from further analysis) if the sigmoid fits are poor, for example with an r-squared (coefficient of determination) value less than 0.5.
In some embodiments, each accuracy value is multiplied by 100, to convert from proportion to percent of the distance to the target. Some embodiments subtract 100 from each accuracy value such that hypometric movements (movements ending short from the target) have negative values and hypermetric (movements ending past the target) have positive values. For example, if movement kinematics are being used, an embodiment may compute the median of each kinematic metric across a plurality of movements, separately for each individual player, and separately for each movement-type (primary versus corrective). Doing so may yield, for example, a total of 10 values for the movement kinematic metrics for each task: 5 kinematic metrics (median speed, MAD precision, median accuracy, median reaction time, and median swipiness) for primary movements and 5 kinematic metrics (median speed, MAD precision, median accuracy, median reaction time, and median swipiness) for corrective movements. Some embodiments compute a z-score (a statistical measurement of distance from a mean or average) for each movement kinematic metric across players. Similar processing may be used for other player data, such as shot performance metrics.
Some embodiments compare the input controller sensitivities (e.g., mouse sensitivities) with the movement kinematic metrics. To mitigate the confounding effects of the sensitivity of the input controller (e.g., mouse sensitivity), some embodiments regress out sensitivity from the movement kinematics values. An embodiment may correlate between mouse sensitivity and each movement kinematic metric for the primary movements, e.g. from a dataset of 32 pro and semi-pro esports players, using, e.g. a Gridshot task and a Sixshot task. An embodiment may use statistical results from Spearman's Rank Correlation (e.g. rho=Spearman's Correlation Coefficient and p=p-value; e.g. using p-value<0.05).
An embodiment may use residual correlations after z-scoring the movement kinematic metrics across players and regressing out mouse sensitivity from the movement kinematic metrics. Mouse sensitivity may be correlated with movement reaction time, movement speed, swipiness and other performance metrics. The correlation may be removed by the regression procedure. Embodiments may use correlations, before and after z-scoring and regressing out mouse sensitivity, respectively, for corrective movements.
An example regression operation includes Equations 11 through to 13 in order. Firstly, the pseudo-inverse of the mouse sensitivity array a# may be calculated and multiplied by the matrix of z-scored kinematic metrics Y to produce the estimate {circumflex over (x)}:
{circumflex over (x)}=a#Y (11)
Subsequently, an embodiment may compute the estimate Ŷ by multiplying {circumflex over (x)} with the z-scored mouse sensitivity array A:
Ŷ=A{circumflex over (x)} (12)
where subtracting Y{circumflex over ( )} from y gives the residuals (E) of the kinematic performance matrix:
E=Y−Ŷ (13)
Flicking-Skill Assessment
Performance, e.g. motor acuity, may be is determined based on one of the speed performance metrics and one of the variability (or precision) performance metrics, optionally with one or more of the pre-processing steps and post-processing steps. Embodiments may measure a flicking skill assessment (FSA) or motor acuity using a speed-accuracy tradeoff. A statistical analysis or graph may compare shot speed as a function of the median shot variability for, e.g., Gridshot and Sixshot tasks. One or more players assessed may follow the expected pattern in behavior: lower precision and faster speed in Gridshot compared to Sixshot. This demonstrates that players respond strategically and appropriately to the different demands of the two tasks. Lines may be plotted each of which connects the multiple (e.g., two) data points from one single player corresponding to the multiple (e.g., two) tasks. After plotting or analysis, the resulting plot or analysis may be re-plotted or analyzed in terms of shot precision (e.g. a measure being 1 over or divided by shot variability). The resulting curves may represent the transformed lines from the first analysis or plot. A diagonal dashed line or other plot may be plotted from the origin (0,0) to three times the standard deviation plus the mean of the distribution of shot precision (x-axis) and shot speed (y-axis) across all players. In some embodiments, the intersection of individual curves with the diagonal may indicate flicking skill. Some embodiments obtain a flicking skill value for a player, by identifying the point along the diagonal line that intersects with that player's speed-accuracy tradeoff curve; in some embodiments arbitrary skill units may be assigned to the relative position of the intersection or distance of the intersection from the origin. This relative skill may be player vs. other player(s); or player vs. performance of that same player at another time. Speed-accuracy tradeoff curves located towards the upper right indicate better performance, e.g., higher precision and faster speed. Notably, as demonstrated elsewhere herein, these flicking skill values are related to certain aspects of movement kinematics; movement kinematics may provide details as to the basis for specific differences in skill between players. A diagonal line drawn through each player's speed-accuracy tradeoff curve may rank each player such that, e.g., the speed-accuracy tradeoff curves further from the origin associated with players having higher measured skill. Such an embodiment may compare motor acuity values across players to determine which player has better skill; over time (for an individual player) to characterize that player's improvement over time; or across conditions, e.g., hardware, software, and various other factors (e.g., mouse, mouse pad, rendering latency, posture, exercise, diet and dietary supplements, sleep, etc.), for a group of players or for an individual player, to determine the impact of these factors on performance (e.g., to show that a new gaming mouse is better than a competing product).
In one embodiment, FSA is computed as the value along a diagonal or other plot or line that intersects a player's speed-accuracy tradeoff curve. There are a variety of ways to compute this flicking skill value. For example, in one embodiment the diagonal line is transformed back to the shot speed vs. shot error plot and the intersection could be found there; the diagonal may be a scale line indicating skill. In one embodiment a mathematical formula is used to find the intersection. In embodiments a line with different slope is used, or the x- and/or y-axes are rescaled.
For example, for each player one line (e.g., a best fit line) may be drawn through the multiple target type graph plots for that player; to get one line per player. The left panel of
The right panel of
Some embodiments may rank flicking skill across multiple players, or may rank an individual player's change in flicking skill over time, for example such that the speed versus precision curves intersect the axis of flicking skill (the dashed diagonal line) more towards the upper right (e.g. as speed and precision increase) as the player improves over time.
In some embodiments, the axis of flicking skill (e.g., indicating as the diagonal dashed line in
The graphs shown in
An embodiment may perform a multiple linear regression (e.g. on a number of data points, each corresponding to a summary (e.g. average or median) of speed in relationship to precision, accuracy or variability). A statistical analysis may determine if there is a change in performance or motor acuity, e.g. if there is a substantial improvement in speed (e.g., greater than 5%) with no substantial change in variability (e.g., no more than 5%), precision or accuracy, or a substantial improvement in variability, precision or accuracy (e.g., greater than 5%) with no substantial change in speed (e.g., no more than 5%), or a substantial improvement in both speed and accuracy (e.g., greater than 5%). Other measures or thresholds may be used to define or deem a change in motor acuity, e.g. a change in motor acuity may occur if there is a >10% speed improvement with no substantial change in variability, or other combinations of changes with different thresholds. A statistical hypothesis test may be used to determine the value or values that specify whether or not there is a substantial change in speed and/or variability, precision or accuracy to determine if there is a statistically significant change in motor acuity, e.g. if there is a statistically significant improvement in speed with no statistically significant evidence for a change in variability, precision or accuracy, or a statistically significant improvement in precision (or variability or accuracy) with no statistically significant evidence for a change in speed, or a statistically significant improvement in both speed and precision (or variability or accuracy). Such an embodiment may determine if there is a difference in motor acuity across players, across conditions, or over time.
In some embodiments, performance metrics other than precision may be compared to or analyzed in relationship to speed, e.g. accuracy or variability of the player's movements or input to a game. The precision, accuracy, variability or other measure of the player analyzed in relation to speed may be, for example, that of player movement kinematics performance metrics, shot performance metrics, or another player performance metric. Other measures of player speed may be used, for example movement speed, shot speed, or other variations of a speed metrics including a target presentation duration threshold as discussed elsewhere herein. Embodiments using precision as a performance metric, and may use measures of precision such as movement precision, shot precision, or other measures. In one embodiment, a comparison of the motor acuity of two players (or to compare a player with themself over time) includes a statistical analysis of whether the speed increases (over time for a single player) or is larger (comparing different players) with no statistically significant evidence for a change in precision (or vice versa).
In some embodiments movement speed or movement variability (e.g. parsed as described herein) may be used as a measure instead of shot speed or shot variability, since in some FPS tasks player movement has a landing point, or ending of the movement. For some applications, there may be no shot, e.g. a task where the player or user is to aim at the center of a target (e.g., move a crosshair to the middle of a circle), with no click or shot input, but movement can also be a measure in applications where there is a shot. In some FPS tasks, a flick may involve move, stop movement, and shooting; while swiping may include shooting while moving, and in some of these tasks movement alone may be a measure of skill. Speed may be for example, how quickly a trigger is pulled, or a peak or average speed of movement, e.g. mouse movement, movement on a mousepad, movement of degrees of rotation, etc.
Some embodiments test the flicking skill values for normality, and some embodiments use non-parametric approaches to statistical analysis, for example, to determine statistically significant differences in flicking skill between players, between groups of players, or over time.
Some embodiments compare the input controller sensitivities (e.g., mouse sensitivities) with the flicking skill values. To mitigate the confounding effects of the sensitivity of the input controller (e.g., mouse sensitivity), some embodiments regress out sensitivity from the flicking skill values. In one example validation or this, correlations between mouse sensitivity and flicking skill for 32 pro and semi-pro esports players were analyzed. Residual correlations after z-scoring the flicking skill values across players was performed, with regressing out mouse sensitivity from the flicking skill values. In this example, mouse sensitivity was correlated with flicking skill and the correlation was removed by the regression procedure. Statistical analysis was performed with Spearman's Rank Correlation: rho=Spearman's Correlation Coefficient and p=p-value; an example p-value being <0.05.
In operation 400, data may be collected from players, e.g. providing input to games or gamified tasks. For example, data from a number of players may be collected to create a database or collection of player performance metrics, such as movement data, shot data, speed data, precision data, or other data. Such a data store may be used to compare a newly seen player to other players. Data from a single player to be evaluated may be collected, e.g. in an embodiment determining improvement of a user over time. In one embodiment, the data for each player is collected for more than one target type.
In operation 410, a player, e.g. a newly seen player or a player having provided input earlier in time, may provide input to a game or gamified task.
In operation 420, performance or motor acuity for the newly seen player may be calculated, evaluated or determined relative to a set of other players or the same player earlier in time (e.g. as collected in operation 400). This may be performed by, for example, determining, for a number of different target types (e.g., different target sizes), speed of the player (e.g. of shooting or movement) and one of precision, accuracy and variability of the player (e.g. of shooting or movement), and determining motor acuity or flicking skill based on the relationship between speed of the player and the one of precision, accuracy and variability of the player. This analysis may be performed by, e.g. statistical analysis, graphing, statistical analogies of graphing, or other methods as described herein. Performance may be calculated by for example, determining a line (e.g. a best-fit line) corresponding to or fitting with plots of a relationship of speed of the player and variability for a plurality of target types (e.g. a different plot for each target size); converting the line to a curve (e.g. by plotting a curve using the inverse of one of the line's parameters); and calculating performance based on the location of the curve in a graph (e.g. the distance of the curve from a graph origin or other point).
In operation 430 the evaluation may be output. The evaluation may be used, for example: to compare a number of players to determine which player has better skill or to compare an individual player over time to characterize that player's improvement over time. For example, the comparison may be used to determine the impact of equipment factors on performance (e.g., to show that a new gaming mouse is better than a competing product). Some embodiments maintain a database of players and their flicking skill assessment scores such that players are notified when they achieve a certain ranking in comparison to other players (e.g., top 10th percentile). Some embodiments display a leaderboard of the top (e.g., 100) players with the highest flicking skill assessment scores. Some embodiments notify esports organizations, leagues, team owners, and coaches about flicking skill assessment rankings from each of a plurality of players.
Other operations or series of operations may be used.
In some embodiments, human participants or players provide input to a movement task (e.g. moving to the center of a circle) or other tasks such as gamified tasks, training, therapy, etc. remotely (e.g. from a server 300 providing a game) using their own gaming set-up such as user computer 200. A gaming set-up such as computer 200 may include for example hardware such as a PC, monitor (e.g. monitor 140 of
While one user computer 200 and one server 300 is shown, other embodiments may use multiple such computers, e.g. connected by a network such as internet 290.
Operating system 115 may be or may include code to perform tasks involving coordination, scheduling, arbitration, or managing operation of computing device 100, for example, scheduling execution of programs. Memory 120 may be or may include, for example, a Random Access Memory (RAM), a read only memory (ROM), a Flash memory, a volatile or non-volatile memory, or other suitable memory units or storage units. Memory 120 may be or may include a plurality of different memory units. Memory 120 may store for example, instructions (e.g. code 125) to carry out a method as disclosed herein, and/or data such as documents, etc.
Executable code 125 may be any application, program, process, task or script. Executable code 125 may be executed by controller 105 possibly under control of operating system 115. For example, executable code 125 may be one or more applications performing methods as disclosed herein. In some embodiments, more than one computing device 100 or components of device 100 may be used. One or more processor(s) 105 may be configured to carry out embodiments of the present invention by for example executing software or code.
Input devices 135 may be or may include game controllers (e.g. a Wii controller), gyroscope sensors, EEG sensors, EMG sensors, a mouse, a keyboard, tablet computing device, a touch screen or pad or any suitable input device or combination of devices. Output devices 140 may include one or more displays, speakers and/or any other suitable output devices or combination of output devices. Any applicable input/output (I/O) devices may be connected to computing device 100, for example, a wired or wireless network interface card (NIC), a modem, printer, a universal serial bus (USB) device or external hard drive may be included in input devices 135 and/or output devices 140.
Embodiments of the invention may include one or more article(s) (e.g. memory 120 or storage 130) such as a computer or processor non-transitory readable medium, or a computer or processor non-transitory storage medium, such as for example a memory, a disk drive, or a USB flash memory, encoding, including or storing instructions, e.g., computer-executable instructions, which, when executed by a processor or controller, carry out methods disclosed herein.
Some embodiments assess the deftness with which a player can adjust their performance in response to task demands. While playing an FPS game, an opponent can appear at a distant location (e.g., rendered as a small target on the player's screen), or at a nearby location (e.g., rendered as a large target on the player's screen). A good player can flexibly shift along their SAT function or line in response to such circumstances, prioritizing speed for nearby opponents (e.g., large targets) and prioritizing precision for distant opponents (e.g., small targets). Some embodiments compare performance when different target sizes are presented in separate rounds of play versus when the different target sizes are interleaved in each round of play. An ideal player will exhibit exactly the same performance in both cases (separate rounds versus interleaved). Most players will, however, exhibit a decrease in performance (e.g. flicking skill and/or movement kinematics) when the different target sizes are interleaved because that version of the task has the additional requirement that players must decide for each target how best to trade off speed versus precision. Hence, the use of a comparison of performance between playing with different target sizes in separate rounds versus interleaved characterizes a player's ability to optimize the trade-off of speed versus precision and improves the technology of performance evaluation. The ratio of FSA when different target sizes are interleaved versus when different target sizes are in separate rounds may be used as a measure of player's ability to optimize the trade-off of speed versus precision. Some embodiments maintain a database of players and their FSA ratio. Some embodiments notify players when they achieve an ranking of their FSA ratio in comparison to other players (e.g., top 10th percentile), or display a leaderboard of the top (e.g., 100) players with the highest FSA ratios, or notify esports organizations, leagues, team owners, and coaches about FSA ratios. Any comparison of FSA when different target sizes are interleaved versus when different target sizes are in separate rounds may be substituted for an FSA ratio.
Some embodiments include a training program wherein an individual participates in a plurality of sessions across a plurality of days, the individual practices a plurality of repeats of each of a plurality of tasks during a training session, wherein each of the different tasks includes targets with different sizes and/or presentation durations, and wherein performance is assessed by flicking skill and/or movement kinematics for each session. Any changes or improvements in FSA may be displayed to the player as a graph or some other visual rendering.
In some embodiments, FSA is used to evaluate the impact of hardware, software, and various other factors (e.g., mouse, mouse pad, rendering latency, posture, exercise, diet and dietary supplements, sleep, etc.) on performance. Embodiments make a comparison of the FSA for a player or players across conditions, thereby providing actionable insights about such factors, for example, to make a choice about which mouse or mouse pad (or other hardware) to purchase or to offer advice about exercise, diet and sleep so as to optimize performance.
Some embodiments may improve rehabilitation technology, and may include gamified rehabilitation, for example, for patients recovering from or receiving therapy for stroke, traumatic brain injury, cerebral palsy, or arm injury. These embodiments may measure improvement in motor acuity over time. Some embodiments may improve concussion assessment, to determine if an individual may have suffered a concussion or brain injury, by comparing their performance with a baseline that was measured before the incident that may have caused an injury. Some embodiments include an assessment to determine if an athlete should return to play following a concussion. Some embodiments include assessment and training for individuals who are learning to use a prosthetic limb. Some embodiments include assessment and training for stroke rehabilitation. Some embodiments include assessment and training for patients with movement disorders (e.g., cerebral palsy). Embodiments may maintain a database of patients and their motor acuity scores over time, that may be displayed to the patients as a graph or some other visual rendering. Embodiments may provide a dashboard for clinicians to track the progress of their patients and to support decisions, e.g., about which interventions may be most effective or most appropriate, decisions about whether an athlete should be sidelined or allowed to play, etc.
The capacity of embodiments of the invention, e.g. using FSA, to provide, over a short period of time with no ceiling or floor effect, a measure of motor acuity independent of strategy and SAT, makes it an improvement over prior technology and tools for a wide range of both basic and applied research and therapies; for example, visuomotor psychophysics and rehabilitation. There is a paucity of academic studies examining motor acuity. Embodiments of the invention may fill this critical gap in the study of human motor behavior and visuomotor psychophysics. As with any other instance of motor skill learning that is characterized by reinforcement learning, physical rehabilitation aims to improve motor acuity. Rehabilitation must also assess performance to provide motivating feedback for the patient and clinically relevant data for the therapist or clinician to guide rehabilitation. Neuroplasticity drives motor learning, which depends on movement repetition and intensity. Neuroplasticity is also facilitated by active task engagement and enjoyment. Furthermore, calibrating task difficulty to an individual's skill level is critical for rehabilitation, because competency is an intrinsic motivator. Embodiments, e.g. an FSA method, may satisfy all of these criteria, and improve such technology, by for example providing a challenging, adaptive, and engaging task based on repetitive movement behavior. Incorporating FSA methods into gamified rehabilitation therapy technology may enable objective quantification of behavioral or motor performance (e.g., kinematics, dynamics), and may be rapidly and inexpensively deployed at scale and remotely provided to large populations.
Some embodiments include gamified assessment of sensorimotor and cognitive fitness. These embodiments can include immersive games for military personnel that execute on mobile devices (e.g. phone, tablet, etc.) and that assess and monitor changes in cognitive fitness, readiness, and performance across the military lifecycle and in the field. In some of these embodiments, baseline performance is measured in an initial session and repeated gameplay monitors changes in cognitive performance relative to an individual's baseline. In some embodiments, the assessment is sensitive to subtle changes in cognitive fitness (e.g., due to fatigue, injury, mental illness, drug or alcohol use, or any of a number of other factors). Embodiments may provide a dashboard for tracking the sensorimotor and cognitive fitness of personnel to support decisions, e.g., about whether intervention is warranted, whether personnel should be temporarily removed from active duty or returned to active duty, etc.
There are many advantages of the disclosed gamified assessments of sensorimotor and cognitive fitness over other approaches including standard neuropsychology assessments. Embodiments of a gamified assessment of sensorimotor and cognitive fitness may improve on assessment technologies, as embodiments may be entertaining, absorbing, objective, and easily distributable because they may not require specialized hardware, peripherals. Embodiments may not require the presence of a medical professional. Embodiments may be individualized (e.g., relative to each individual's baseline) and hence be sensitive to subtle changes in cognitive fitness that may be undetectable with a standard neuropsych battery. Embodiments may take less time to complete than a standard neuropsych battery.
Some embodiments detect, e.g., using a computer system as described herein, sandbagging (purposefully performing poorly on a baseline assessment) by analyzing performance variability. Human performance exhibits stereotypical error distributions. Although a player may attempt to sandbag (e.g., perform worse than they are capable of), they are unlikely to be able to mimic the stereotypical error distributions while sandbagging. In addition, players of lower skill will have less fast and more variable (or less precise or less accurate) responses overall but still follow speed-accuracy tradeoff incentives of the task, whereas players who intentionally sandbag would have poor speed and precision (or more variability or less accuracy) regardless of task incentives. Embodiments may notify coaches, commanding officers, managers, etc. when sandbagging is detected.
In one validation of some embodiments, performance data from 32 professional and semi-professional male esports players (mean age=22.47±3.62) were collected and then analyzed. The players specialized in different game titles: 4 for a Valorant game, 10 for a PUBG game, and 18 for a Rainbow Six Siege game. Each participant played several rounds of a Gridshot task and several rounds of a Sixshot task. It was predicted that players would exhibit systematic differences in movement behavior between Gridshot and Sixshot tasks. Gridshot's large target size incentivizes players to maximize score by moving as quickly as possible with relatively low shot precision. Sixshot targets are much smaller and require high precision, which consequently means that players are incentivized to slow down.
In one demonstration and validation of FSA, data was analyzed from a large number of amateur players, each of whom played multiple rounds of the Adaptive Reflexshot task with 3 different target sizes. Each round of play had a single target size and target presentation duration was titrated throughout each round of play based on the player's performance. Different target sizes were presented during separate rounds of play.
Some embodiments compare flicking skill across a number of players.
Embodiments may evaluate flicking skill and movement frequency. Human flicking performance typically exhibits variability and error. At times, a primary movement will have very little error, while others will require subsequent corrective movements to successfully land on and shoot a target. For large targets, it is expected that the proportion of targets destroyed which required a corrective movement would be fewer. This is because even with movement error, primary movements are more likely to land on the target. For small targets, on the other hand, players will often have to make corrective movements to accommodate for the targets being much smaller. In one example validation, data from pro (and semi-pro) esports players was used for validation. In this example validation the percentage targets that were destroyed with a single (primary) movement or with a pair of movements (primary and corrective), for each of two tasks, was used. A statistical analysis (e.g. using error bars) produced the standard deviation across players. The results reflected the predicted pattern, with the number of single-movements to destroy a target being greater in Gridshot tasks than in Sixshot tasks across all players. Similarly across all players, single-movements to destroy the target were less frequent than double-movements to destroy the target in Sixshot tasks. Furthermore, with the exception of two players, the opposite was true in Gridshot tasks; where the majority of targets destroyed was achieved following a single movement.
In one example validation for FSA, flicking skill is predictive of the number of movements required to destroy targets. Flicking skill may be plotted on an x-axis (e.g., after regressing out mouse sensitivity) and percentage of targets destroyed following a single-movement may be plotted on a y-axis for a Gridshot task and for a Sixshot task Such validation may use statistical results from Spearman's Rank Correlation: rho=Spearman's Correlation Coefficient and p=p-value; e.g. p-value<0.05. Flicking skill may in one validation be positively correlated with a proportion of targets that were destroyed with a single movement [p=4.19e-04] in Gridshot. This suggests that players with better flicking skill were more likely to require only a single movement to destroy a target, and therefore were more efficient with their movements in Gridshot.
Movement kinematics may be task dependent.
Unless explicitly stated, the method embodiments described herein are not constrained to a particular order or sequence. Furthermore, all formulas described herein are intended as examples only and other or different formulas may be used. Additionally, some of the described method embodiments or elements thereof may occur or be performed at the same point in time.
While certain features of the invention have been illustrated and described herein, many modifications, substitutions, changes, and equivalents may occur to those skilled in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.
Various embodiments have been presented. Each of these embodiments may of course include features from other embodiments presented, and embodiments not specifically described may include various features described herein.
The present application is a continuation-in-part of prior U.S. application Ser. No. 18/129,476, filed on Mar. 31, 2023, which in turn claims benefit from U.S. provisional Patent Application No. 63/335,479, filed on Apr. 27, 2022, entitled “FLICKING SKILL ASSESMENT”, and U.S. provisional Patent Application No. 63/357,194, filed on Jun. 30, 2022, entitled “FLICKING SKILL ASSESMENT”; the present application is also a continuation-in-part of prior U.S. application Ser. No. 17/975,769 filed on Oct. 28, 2022, which in turn is a continuation of prior U.S. application Ser. No. 16/121,210, filed on Sep. 4, 2018, entitled “SENSORIMOTOR ASSESSMENT AND TRAINING”, which in turn claims the benefit of prior U.S. provisional application 62/554,212, filed on Sep. 5, 2017, each incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 63335479 | Apr 2022 | US | |
| 63357194 | Jun 2022 | US | |
| 62554212 | Sep 2017 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16121210 | Sep 2018 | US |
| Child | 17975769 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 18129476 | Mar 2023 | US |
| Child | 18360915 | US | |
| Parent | 17975769 | Oct 2022 | US |
| Child | 18129476 | US |
