WO2015076682A1 - Système et procédé pour estimer ou prédire un résultat d'un match dans un événement sportif - Google Patents

Système et procédé pour estimer ou prédire un résultat d'un match dans un événement sportif Download PDF

Info

Publication number
WO2015076682A1
WO2015076682A1 PCT/NZ2014/050012 NZ2014050012W WO2015076682A1 WO 2015076682 A1 WO2015076682 A1 WO 2015076682A1 NZ 2014050012 W NZ2014050012 W NZ 2014050012W WO 2015076682 A1 WO2015076682 A1 WO 2015076682A1
Authority
WO
WIPO (PCT)
Prior art keywords
sporting event
team
outcome
context
determining
Prior art date
Application number
PCT/NZ2014/050012
Other languages
English (en)
Inventor
Paul John BRACEWELL
Original Assignee
Crichq Limited
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Crichq Limited filed Critical Crichq Limited
Publication of WO2015076682A1 publication Critical patent/WO2015076682A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"

Definitions

  • the present invention generally relates to a system and method for assessing or predicting a match outcome in a sporting event.
  • Embodiments of the present invention relate to a contextual system for real-time prediction of match outcomes in a sporting event.
  • Embodiments of the present invention relate to a system for determining a factor which correlates the playing conditions to the match outcome and a system for determining a measure of a player's impact on the match outcome.
  • Examples of existing systems for predicting an outcome in a sporting match include systems based on the Duckworth & Lewis method (D/L method).
  • D/L method Duckworth & Lewis method
  • a system using this method is used in cricket to forecast scores in limited-overs cricket matches.
  • the D/L method is a mathematical formula designed to calculate a target score for a team batting second in a limited-overs match interrupted by weather or other circumstances.
  • a system for predicting the outcome of a cricket match, using the D/L method is restricted to rain affected games, second innings forecasts, and limited-overs matches. In these games where play is interrupted, the D/L method is applied to predict the score if the game was not interrupted and the play continued.
  • match prediction systems for cricket using other algorithms such as the WASP and Crampton algorithms, have similar restrictions.
  • the WASP algorithm is restricted to limited-overs matches and can only provide a score projection in the first innings and only a probability of winning in the second innings.
  • the behavioural context involves data representing previous events that occurred in a sporting event. Examples of previous events include points scored over an amount of resource consumed to score those points.
  • the present invention broadly consists in a computer-implemented method for determining an outcome of a sporting event, the method comprising :
  • the contextual engine and prediction engine comprise one or more computing devices.
  • the computing device(s) of the contextual engine is/are the same as the computing device(s) of the prediction engine.
  • the computing device of the contextual engine(s) is/are separate from the computing device(s) of the prediction engine.
  • the outcome includes a result of the sporting event.
  • the result comprises a win, a loss, or a draw.
  • the sporting event comprises team sporting events or individual sporting events.
  • the sporting event comprises cricket, rugby Union, Rugby League, American football, Australian Rules football, netball, basketball, soccer, hockey, tennis, squash, badminton, and/or ice-hockey.
  • the method comprises determining the outcome of one or more formats of the sporting event.
  • the formats may be limited-over matches where each team bats once (such as T20 cricket, 40 overs cricket, and/or 50 overs cricket for example) and/or multiple-innings matches where each team can bat multiple times per match (such as first class, test, two-day, and/or three-day matches for example).
  • the method comprises determining the outcome at a start of the sporting event.
  • the start of the sporting event is when the first ball in an innings is bowled.
  • the start of the sporting event is a first second of the sporting event.
  • the method comprises determining a probability of a team (or side) winning. In an embodiment, the method comprises determining the probability of the team winning before the sporting event starts. In an embodiment, the method comprises determining the probability of a first team beating a second team. In an embodiment, the method comprises determining the probability of the first team beating the second team based at least in part on a score in a previous sporting event between the first team and a third team and on a score in a previous sporting event between the second team and the third team.
  • the method comprises determining a score prediction of the sporting event. In an embodiment, the method comprises determining the score prediction before the sporting event starts. In an embodiment, the score prediction is based on one or more contextual factors. In an embodiment, the contextual factors are quantifiable. In an embodiment, the contextual factor comprises one or more of weather conditions, pitch conditions, size-of-ground, and out-field conditions.
  • the method comprises determining a behavioural context of the sporting event, wherein the outcome is determined based at least partly on the behavioural context.
  • the momentum factor is based at least partly on the behavioural context.
  • the behavioural context comprises information relating to previous events that occurred in the sporting event.
  • the method comprises providing the behavioural context from a data capture system.
  • the behavioural context comprises points scored over an amount of a resource consumed to score those points, wherein a resource is consumed by a team in the sporting event to score points.
  • the behavioural context comprises in-game scoring rate.
  • the in-game scoring rate comprises a run-rate or wicket-loss rate over a previous number of overs in the sporting event.
  • the method comprises determining the in-game scoring rate over two or more overs in the sporting event.
  • the method comprises determining the in-game scoring rate over five or more overs in the sporting event.
  • the momentum factor is based on recent changes in score.
  • the momentum factor is based on changes in resources.
  • the sporting event is cricket, the resources include batsmen and the changes in resources relate to batsmen dismissed.
  • the method comprises determining a historical context of the sporting event, wherein the outcome is determined based at least partly on the historical context.
  • the historical context comprises information relating to previous sporting events.
  • the historical context for a team comprises a historical scoring rate of the team.
  • the method comprises determining a current context of the sporting event, wherein the outcome is determined based at least partly on the current context.
  • the current context comprises information relating to a current state of the sporting event.
  • the method comprises providing the current context from a data capture system.
  • the current context comprises information on available resource(s) and/or a current difference between scores of teams in the sporting event.
  • the information relating to available resource(s) comprises overs remaining or wickets remaining.
  • the method comprises applying predictive algorithms to determine the outcome of the sporting event.
  • the present invention broadly consists in a method for displaying an outcome of a sporting event, wherein the outcome is determined by the method of the first aspect described above.
  • the present invention broadly consists in a system for determining an outcome of a sporting event, the system comprising : a contextual engine for determining a momentum factor at any point in time of the sporting event, wherein the momentum factor is based on previous events in the sporting event; and
  • a prediction engine for determining the outcome of the sporting event based on at least the momentum factor.
  • the outcome includes a result of the sporting event.
  • the result comprises a win, a loss, or a draw.
  • the sporting event comprises team sporting events or individual sporting events.
  • the sporting event comprises cricket, rugby Union, Rugby League, American football, Australian Rules football, netball, basketball, soccer, hockey, tennis, squash, badminton, and/or ice-hockey.
  • the method comprises determining the outcome of one or more formats of the sporting event.
  • the formats may be limited-over matches where each team bats once (such as T20 cricket, 40 overs cricket, and/or 50 overs cricket for example) and/or multiple-innings matches where each team can bat multiple times per match (such as first class, test, two-day, and/or three-day matches for example).
  • the system comprises an initial prediction engine for determining the outcome at a start of the sporting event.
  • the start of the sporting event is when the first ball in an innings is bowled.
  • the start of the sporting event is a first second of the sporting event.
  • the initial prediction engine is configured to determine a probability of a team (or side) winning. In an embodiment, the initial prediction engine is configured to determine the probability of the team winning before the sporting event starts. In an embodiment, the initial prediction engine is configured to determine the probability of a first team beating a second team. In an embodiment, the initial prediction engine is configured to determine the probability of the first team beating the second team based at least in part on a score in a previous sporting event between the first team and a third team and on a score in a previous sporting event between the second team and the third team.
  • the initial prediction engine is configured to determine a score prediction of the sporting event. In an embodiment, the initial prediction engine is configured to determine the score prediction before the sporting event starts. In an embodiment, the score prediction is based on one or more contextual factors. In an embodiment, the contextual factors are quantifiable. In an embodiment, the contextual factor comprises one or more of weather conditions, pitch conditions, size-of-ground, and out-field conditions. In an embodiment, the system comprises a contextual engine for determining a behavioural context of the sporting event, wherein the outcome is determined based at least partly on the behavioural context. In an embodiment, the momentum factor is based at least partly on the behavioural context. In an embodiment, the behavioural context comprises information relating to previous events that occurred in the sporting event.
  • the behavioural context comprises points scored over an amount of a resource consumed to score those points, wherein a resource is consumed by a team in the sporting event to score points.
  • the behavioural context comprises in-game scoring rate.
  • the in-game scoring rate comprises a run-rate or wicket-loss rate over a previous number of overs in the sporting event.
  • the method comprises determining the in-game scoring rate over two or more overs in the sporting event.
  • the method comprises determining the in-game scoring rate over five or more overs in the sporting event.
  • the momentum factor is based on recent changes in score.
  • the momentum factor is based on changes in resources. In one
  • the resources include batsmen and the changes in resources relates to batsmen dismissed.
  • the system is configured to determine a historical context of the sporting event, wherein the outcome is determined based at least partly on the historical context.
  • the historical context comprises information relating to previous sporting events.
  • the historical context for a team comprises a historical scoring rate of the team.
  • the system is configured to determine a current context of the sporting event, wherein the outcome is determined based at least partly on the current context.
  • the current context comprises information relating to a current state of the sporting event.
  • the current context comprises information on available resource(s) and/or a current difference between scores of teams in the sporting event.
  • the information relating to available resource(s) comprises overs remaining or wickets remaining.
  • the system comprises a data capture system.
  • the data capture system is configured to provide the behavioural context to the contextual engine.
  • the data capture system is configured to provide the current context to the contextual engine.
  • the system comprises a data processing engine for receiving data from, the contextual engine.
  • the data processing engine is configured to collate, combine, manipulate and transform data.
  • the data processing engine is configured to convert data to a flat file structure.
  • the system comprises a prediction engine for applying predictive algorithms to determine the outcome of the sporting event.
  • the system comprises a delivery interface for transmitting match outcome to relevant parties.
  • the delivery interface is further configured to feedback the match outcome to the initial prediction engine.
  • the present invention broadly consists in a computer-implemented method for correlating a performance of a player or team to a playing condition in a sporting event, the method comprising :
  • the computing device for determining the actual match outcome is the same as the computing device for comparing the actual match outcome of the predicted match outcome. In an alternative embodiment, the computing device for determining the actual match outcome is separate from the computing device for comparing the actual match outcome of the predicted match outcome.
  • the method comprises determining a match outcome in a sporting event based at least in part on the correlation factor.
  • the predicted scoring rate is determined according to the method of the first aspect or system of the third aspect, and respective embodiments of the method and system, described above.
  • the first playing condition comprises the pitch conditions
  • environment conditions temperature or humidity
  • outfield conditions condition of the ball
  • lighting conditions and/or whether the team is playing on home ground or away.
  • the correlation factor is a pitch correlation factor, wherein the pitch correlation factor puts the score of a player or team on a particular pitch into context and can be used to indicate a player or team's ratings
  • the method comprises determining the correlation factor based on the scoring rate of both sides, and the peak, relative and contextual run rates over specified periods of time.
  • the method comprises determining an overall weighting to estimate resources used and resources remaining.
  • the resources comprise time, overs and/or balls.
  • the method comprises determining the correlation factor based on minimum information required to record the result of a game. In an embodiment, the method comprises determining a pitch correlation factor for cricket, wherein pitch correlation factor is based on one or more of the attributes listed below:
  • the present invention broadly consists in a computing device configured to perform the method of the fourth aspect, and any respective embodiment of the fourth aspect, described above.
  • This invention may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, and any or all combinations of any two or more said parts, elements or features, and where specific integers are mentioned herein which have known equivalents in the art to which this invention relates, such known equivalents are deemed to be incorporated herein as if individually set forth.
  • '(s)' following a noun means the plural and/or singular forms of the noun.
  • 'and/or' means 'and' or 'or' or both.
  • Figure 1 shows an overview of the contextual system according to an embodiment of the present invention for real-time prediction of match outcomes in sport
  • Figures 2 and 2A show a sample output from a data input of Figure 1;
  • Figures 3 and 3A show a sample output from a data processing engine of Figure i;
  • Figure 4 shows application of the system according to an embodiment of the present invention to Rugby Union
  • Figure 5 shows another application of the system according to an embodiment of the present invention to Rugby Union
  • Figure 6 shows application of the system according to an embodiment of the present invention to Short Form Cricket
  • Figure 7 shows another application of the system according to an embodiment of the present invention to Short Form Cricket
  • Figure 8 shows an application of the system according to an embodiment of the present invention to Multiple Innings Cricket
  • Figure 9 shows an application of the system according to an embodiment of the present invention to Multiple Innings Cricket.
  • Figure 10 shows a flow chart for determining a correlation/correction factor between a match outcome and a playing condition according to an embodiment of the present invention .
  • a system of an embodiment of the present invention relates to a data-driven contextual system that determines real-time (or near real-time) predictions of a match outcome within a sporting event, that can be either delivered during or after that sporting event.
  • a 'match outcome' includes one or more of the result (win, loss, or draw for example), the score (points or runs for example), and the margin of victory (the score difference between the teams) of the match.
  • This system is applicable to a variety of team and/or individual sporting events.
  • the system of the present invention may be used for, but not restricted to, cricket, rugby Union, Rugby League, American football, Australian Rules football, netball, basketball, soccer, hockey, tennis, squash, badminton, and/or ice-hockey for example.
  • the system is also applicable to all, or at least a variety of, formats of these types of organised sporting events,
  • this system can be configured to predict a match outcome which includes projected runs scored, margin of victory and probability of winning for all forms of cricket such as limited over matches (for example, T20 cricket, 40 overs, and 50 overs, where each team bats only once per match) as well as two innings matches (such as first class, test, two-day, and three-day matches, where each team can bat up to twice per match).
  • both projected scores and probability of winning estimates are generated from the very start of the match through until completion (first ball of first innings, or first second of time bound event).
  • the system is based on a contextual system for real-time prediction of match outcomes in a sporting event, and delivery of the predictions to interested parties (such as media, coaches, players, administrators, advertisers and spectators) that are meaningful and quantitative indications to the progress of a sporting context.
  • the system is further configured to indicate the probability of each team (or side) in the sporting event winning at any point in time during the match, and also what the likely score for each team is likely to be.
  • the system can calculate the probability of each team winning and the predicted margin of victory (ie. wickets remaining, balls to spare, or runs margin where the sporting event is cricket).
  • the system can make predictions during the first and second innings. Further, based on the progression of a team's score in the current sporting event (ie. the run rate or the wicket loss rate for cricket), the system is configured to calculate a 'momentum' factor (or behavioural context or game pace) for that team (for example, relative recent scoring rate in rugby or when the net difference between scores is less than one scoring play). The system can take this 'momentum' factor into account when making predictions.
  • the method, and system implementing the method, for determining an outcome of a sporting event comprises determining the momentum factor at any point in time of the sporting event and determining the outcome of the sporting event based on at least the momentum factor.
  • the momentum factor is based on the behavioural context (or previous events) in the sporting event.
  • the behavioural context comprises information relating to previous events that occurred in the sporting event (such as points scored over an amount of a resource consumed to score those points, wherein a resource is consumed by a team in the sporting event to score points).
  • the behavioural context comprises in-game scoring rate, such as a run-rate or wicket-loss rate over a previous number of overs in the sporting event (for example, two or more overs).
  • the contextual system 100 for real-time (or near real-time) prediction of a match outcome in a sporting event (which can be delivered during or post event) has six components, which are identified below:
  • Delivery Interface 160 Figure 1 shows how the six components 110-160 of the system 100 interact together. These components will be explained in further detail below.
  • Initial Prediction Engine 110 This initial prediction engine 110 is configured to determine the probability of winning. For example, in an embodiment this initial prediction engine 110 is configured to estimate the likelihood of Team A beating Team B. This estimation may be based on work similar to that described by, and referenced in, Bracewell, P.J., et al. 'Determining the Evenness of Domestic Sporting Competition Using a Generic Rating Engine', Journal of Quantitative Analysis in Sports (2009), Vol. 5, No. 1, pp 1-25.
  • the initial prediction engine 110 may also be configured to make a score prediction.
  • the prediction may be based on quantifiable contextual factors that influence scoring rate (e.g. weather, pitch, size of ground).
  • the initial prediction engine 110 is configured to make predictions prior to the start of the sporting event.
  • the Initial prediction engine 110 Is static during an event and Is updated at the conclusion of an event.
  • the initial prediction system 110 determines the probability of Team A beating Team C at the start of the game based on historical results, such as the outcome of events between Team A and Team B, and between Team B and Team C for example.
  • Data input system 120
  • the data capture or data input system 120 is an electronic or digital system for capturing data regarding the event.
  • the data is then automatically passed into the Contextual Engine via digital/electronic means.
  • the input system is configured for real-time (or near real-time) data input. The result is that data is sent to the contextual engine in real time (or near real-time).
  • the data input system 120 comprises intelligent data processing to minimise data capture to enable access to all levels of a sport without bespoke data entry.
  • the system comprises a timer to track the resource use/consumption (time) within the game and a mechanism for entering the number of points awarded to each team at any point in time (ie. 5 points for a try in rugby union).
  • a system such as CricHQ ( w w . C r i c H 0. :o m ) is suitable for capturing the minimum data elements relative to the resource use/consumption.
  • the data input system transfers the information to a database in real-time (or near real time, or in batch mode for simulation purposes).
  • the database can be accessed by the relevant scoring and contextual engines in (or near real time, or in batch mode for simulation purposes).
  • the cricket scorer enters data into the CricHQ Scoring Application via an internet connected device.
  • internet connected devices include tablets, laptops and smart phones. Scoring of cricket is mandated by the laws of the game.
  • the CricHQ system is configured to score virtually any level of cricket, from international though to children's cricket.
  • the data is preferably extracted and uploaded in a batch mode. In this case score projections and the probability of winning is only relevant for post-match analysis.
  • Figures 2 and 2A show a sample output from the data input system 120.
  • Contextual engine 130
  • the contextual engine 130 comprises:
  • Behavioural context sub-model 132 considers what has happened in the current sporting event prior to current action or time.
  • Historical context sub-model 134 considers what has happened in sporting events completed prior to this sporting event.
  • the data capture system 120 provides data to the behavioural context and current context sub-models 132, 136.
  • the current context sub-model 136 considers the current state of the sporting event at the present play/action or time.
  • the information from the historical context sub-model 134 relates to previous
  • the historical context sub-model 134 is different from the initial prediction engine 110.
  • the initial prediction engine 110 is configured to predict the match outcome prior to the sporting event.
  • the historical context sub-model considers how events and strings of events within previous games have impacted on historical games in progress to then reweight these attributes to recalculate the forecasted outcome.
  • the behavioural context may include in-game scoring rate (run rate or wicket-loss rate over the previous number of overs in the current sporting event), the historical context may include historical scoring rate of the team, and the current context may include the remaining resources available (overs and/or time and wickets remaining) and the required run rate to reach the opposition total.
  • a points/score rate (such as a 'run rate') is the points scored over the amount of resources consumed to score those points.
  • a 'run rate' is the number of runs over the number of balls/overs to achieve those runs.
  • the behavioural context in rugby may include in-game scoring rate (points scoring rate or try scoring rate over the number of minutes lapsed in the current sporting event), the historical context may include historical scoring rate of the team, and the current context may include the remaining resources available (time remaining) and the required scoring rate to reach the opposition total (that is a rugby team may need to outscore the opposition by a factor of 2-to-l to take the lead immediately prior to the completion of the win.
  • the behavioural context is dynamic and depends on the current context of the sporting event.
  • the value of the behavioural context factor may change throughout the sporting event.
  • the value of the behavioural context factor may be constant through the sporting event, or throughout at least a portion of the event.
  • the behavioural context reflects the impact of recent events in this sporting event on the prediction at the current action or time.
  • the contextual engine 130 encompasses all match formats, all innings, and is not restricted to rain affected matches.
  • the contextual engine 130 incorporates match context into the estimates, which can be obtained from all sports types. Examples of match context include:
  • scoring system used for example, the scoring system used in rugby Union
  • cricket may alternatively or additionally incorporate the following types of contextual information (which can be derived from a minimal data capture system in-game):
  • the data processing engine 140 collates, combines, manipulates and transforms data so that the data can be presented to the prediction engine 150.
  • the data processing engine 140 converts data from a transaction (or long-thin) file format to a flat file (or short-fat) structure to enable the calculations to be performed for the relevant time period (or event of interest). Processing of previous events and/or strings or previous events is also processed in this step (for example, time lag or series of time lagged data). Any necessary processing to ensure assumptions of modelling techniques are also performed (for example, square root transformation of count data to stabilise variance and/or induce an approximately normal distribution), Where data is obtained from multiple data sets, these are combined into a single data set (for example, historical individual strike rate combined with strike observed in current match).
  • the first component modelled by the data processing engine 140, and applied in the contextual system is the predicted score.
  • the output from the data processing engine 140 is used in the prediction engine 150.
  • the forecasted score utilises the concept of resources. Simply, resources translate to the scoring or points system of the sporting event.
  • the relationship used to create the model structure is defined below:
  • P(S C ) is the predicted final score for the i th team observed at point c in the
  • c either represents time or some other measure of match prog ress (such as overs or balls bowled) .
  • the i th team represents either one of the two teams competing in the sporting event.
  • S c ,i is the current score for the i th team observed at point c in the sporting event, where c is defined above.
  • R C// is the estimated resources currently consumed by the * team observed at point c in the sporting event, where c is defined above.
  • Ti is the final score achieved by the i h team.
  • the proportion, S C /T / is then used as the dependent variable in a statistical model F[.] where the relevant contextual inputs, Context C/ i, based on the type of sporting event are used as independent variables.
  • the second component modelled is the match result (win, loss, or draw).
  • the outcome is simplified to binary (win, loss) with the draw (specifically for cricket in two innings matches) absorbed into the match prediction via the context engine (in which case the probabilities of winning for each of the team batting first and the team batting second will sum to less than or equal to 1).
  • the structure of this attribute is shown below:
  • P C)i is the probability of winning for the i th team observed at point c in the sporting event, with c defined previously in this section. In defining the model, this outcome is the dependent variable and is either 1 (Team A wins) or 0 (Team B wins).
  • Figures 3 and 3A show a sample output produced by the data processing engine.
  • Table 3 shows an example of the historical context for a batting team.
  • Table 3 Example of historical context for a batting team Prediction Engine 150
  • the prediction engine 150 applies a suite of predictive algorithms (derived using statistical/mathematical/econometric modelling applied to historical sporting event data) to estimate:
  • the estimated margin of victory can be determined from the estimated score.
  • the prediction engine 150 feeds predicted outcomes back to the data processing engine 140.
  • the prediction engine 150 comprises a score prediction sub-model 152 for predicting a score and margin of victory, and an outcome prediction sub-model 154 for predicting the probability of winning.
  • the prediction engine 150 considers the impact of events that occurs in-game. For example, the prediction engine 150 considers recent changes in score, or recent, dramatic changes in resources (batsmen dismissed where the sporting event is cricket). This is where the concept of current match context comes into play, referred to as 'momentum'.
  • the system takes into account the culmination of recent events and the localised impact that the culmination of events has on an outcome.
  • the prediction engine 150 learns from the data that is provided in an automated scoring system, which is built for the purpose of capturing the minimum required data to determine an outcome in the event and competition (for example, runs and wickets in cricket; points scored and type of points scored in rugby; and goals in soccer),
  • the system of the present invention addresses this problem by modelling resources consumed using techniques that impose situational context onto the estimates which is then reflected in the resultant scoring projections (ie. on a low scoring pitch in cricket, the system scales the estimates effectively). Further, strings of recent events also have a statistically significant impact. Accordingly, the dynamic context of a match (recent scoring rate, recent resource change, relative scoring rates) is applied to refine resource estimates.
  • the resources are modelled to account automatically for environmental factors.
  • the prediction engine 150 also defines a margin of victory.
  • the challenge in coming up with a margin of victory is that in some sports (such as cricket), events are loosely time bound. For example, in a 2 innings game - a team could bat as long as they want and are able. Additionally, the margin of victory must also consider the match
  • the prediction engine 150 looks at consecutive strings of recent events or states to overcome the challenge of access to data and designed to create a truly universal system that can be applied to different formats of multiple sporting events. Game information is updated. Game information includes score, calculations performed for the projected score, and calculations performed for the probability of winning. Delivery Interface 160
  • the delivery interface 160 comprises a delivery sub-model 162 for delivering data to interested parties or other devices, and a post-match delivery sub-model for delivering data back to the initial prediction engine 110.
  • Information in the form of data is preferably available for viewing by all interested parties with an internet connected device.
  • An internet connected device includes a tablet, laptop and smart phone. This information is preferably available to parties in a form that does not require them to capture any data not mandated by the laws of cricket for scoring.
  • the delivery sub-model 162 of the delivery interface 160 transfers the data to interested parties and connected d igital/electronic systems automatically.
  • Figures 2 to 7 show examples of the information from the system 100 displayed on user devices.
  • Figure 4 shows application of the system to rugby Union. This figure shows the probability of winning projection (New Zealand ⁇ Team A ⁇ vs France ⁇ Team B ⁇ , Pool Play, Rugby World Cup 2011) .
  • Figure 5 shows another application of the system to rugby Union. This figure shows the score projections (New Zealand vs South Africa, 12 September 2009).
  • Figure 6 shows application of the system to Short Form Cricket (e.g. T20 or 50 Overs). This figure shows the probability of winning projection (New Zealand ⁇ Chasing Team ⁇ vs England ⁇ Setting Team ⁇ , T20, 9 February 2013).
  • Figure 7 shows another application of the system to Short Form Cricket (e.g . T20 or 50 Overs). This figure shows the score projections (New Zealand ⁇ Chasing Team ⁇ vs England ⁇ Setting Team ⁇ , T20, 9 February 2013).
  • Figure 8 shows an application to Multiple Innings Cricket (e.g. First Class, Test or Club Two-day). This figure shows the probability of winning projection (with projected scores overlaid).
  • Figure 9 shows an application to Multiple Innings Cricket (e.g. First Class, Test or Club Two-day). This figure shows the score projections (Sri Lanka vs Pakistan, 1 st Test, 4-7 July 2009).
  • the post-match delivery sub-model 164 delivers final outcomes to the initial prediction engine 110 to update the relevant estimates.
  • Contextual information about the playing conditions is fed into this system in the form of a correlation/correction factor which scales the scores based on the
  • This aspect of the system generates a correlation (or correction) factor that correlates the performance of a player or team to the playing conditions.
  • playing conditions include the pitch conditions, environment conditions (temperature or humidity), outfield conditions, condition of the ball, lighting conditions, and whether the team is playing on home ground or away, for example.
  • the correlation factor may be used as a contextual input for the contextual engine 130 of the system 100.
  • Figure 10 shows a flow chart 1000 which indicates at a high level the flow to derive a context, which may be used in the contextual engine 130.
  • the shaded boxes indicate where data is entered into the system. Shaded boxes with the prefix “model” refers to historical context.
  • the unshaded boxes refer to the creation of current and behavioural contextual attributes which are derived in game.
  • Basic scorecard information is extracted 1002 from the input system, describing the events.
  • the export_ball_by_ball_data.txt file is imported 1004 into the system.
  • the system prepares 1006 the ball by ball data so it can be merged with information about the individuals involved.
  • Basic scorecard information is extracted 1008 from the input system, describing who was involved in the events and a summary of their skill sets.
  • the system imports 1010 the export_team_list.txt file into the system. Data relating to the individuals involved in the event is the processed 1012.
  • the system combines 1014 ball by ball information with details of the individuals involved in each interaction. The ball by ball information is consolidated 1016 and candidates identified for post-match processing.
  • Time dependent variables are calculated 1018 based on the state of the match, including the identification of consecutive events.
  • the system quantifies 1020 the outcome of the event for post match processing.
  • the system prepares 1022 the data set which contains all input variables necessary to calculate probability of winning and projected total for an innings. This includes the calculation of variables that consider relative run rates. These run rates include past run rates, current run rates and required run rates.
  • the system derives 1024 contextual weightings for variables based on the current state of the match.
  • the system accepts 1026 as input a data set which contains the weights necessary to perform the calculations on the input variables derived in the first innings.
  • the system prepares a data set 1028 which contains both the Input variables and corresponding weights to enable calculations to be performed on first innings data.
  • the system prepares 1030 a data set which contains both the input variables and corresponding weights to enable calculations to be performed on second innings data.
  • the data set is ordered 1032 to created sorted data sets.
  • the system extracts 1034 the last probability of winning from the completion of the 1st innings to adjust the early estimates of winning in the 2nd innings. Projections are collated 1036 from first and second innings, depending on the current state of progress of the game being processed. Calculations to ensure continuity in the probability of winning between the end of the first innings and start of second innings are performed in this stage.
  • the system takes as input a data set 1038 that contains the weights necessary to perform the calculations on the input variables derived in the second innings.
  • a data set is prepared 1040 that has the ball-by-ball projected totals for the game in progress.
  • the system prepares 1042 a data set which has the ball-by-ball projected probability of winning for the game in progress.
  • a final data set is prepared 1044 that contains all scored information per ball.
  • Accepted as input is 1046 a data set to ensure data is appropriately formatted.
  • An example of formatting comprises one observation per legal delivery, up the maximum number of legal deliveries to ensure consistency in graphical displays of the preditictions.
  • the correlation factor can be determined based on the scoring rate of both sides, and the peak, relative and contextual run rates over specified periods of time (depending if the batting team is setting a total or chasing a total).
  • a pitch correlation factor puts the score of a player or team on a particular pitch into context and can be used to indicate a player's ratings. For example, a player's rating based on a score obtained from a bowler-friendly pitch may be worth more compared to the same score obtained from a batsmen-friendly pitch.
  • the correlation factor makes use of the bare minimum required to record the result of a game in the context of a match and/or competition. For example, in determining a pitch correlation factor for cricket, the following attributes are derived and used to create the pitch correlation metric (note, these metrics are examples only) :
  • scoring rates can be derived for any team sport and the above list serves as an indication only.
  • the correlation factor is used to scale scores based on playing conditions to allow for comparisons of teams' performances across different playing conditions.
  • the system infers this from the match data, which allows the system to be used at all
  • the match outcome from other players or teams in a sporting event under a particular playing condition defines a correlation factor (or base condition), that can be used as a measure to determine the match outcome of other players or teams in future sporting events under the same or similar playing conditions. .
  • the system uses the minimum statistics collected for all forms of the game, which makes it a universal system applicable to all levels of a sport. With respect to cricket, these conditions have a massive impact upon a potential score, which may be in the order of hundreds. Using historical data and additional variables, hones in on the average case, which omits the nuances of the game, which the simple, contextual system of the present invention captures to give more precise, natural, intelligent estimates. Measure of player impact
  • the system also provides a measure of predicted impact of a player to the overall performance of the team.
  • the system takes into account the change in outcome of the sporting event before and after the player comes onto the pitch or court.
  • the observed outcome on the probability of winning before and directly after a specific event is captured.
  • a specific event for example a delivery in cricket
  • the changes in probability are added together to give an overall indication of match impact.
  • Ball 1 no run is scored resulting in the probability of winning dropping from 0.4305 to 0.4295, meaning a -0.001 impact on winning.
  • Ball 2 results in 3 three runs and a 0.002 increase in the probability of winning.
  • Ball 3 is hit for a boundary (4 runs), increasing the probability of winning by 0.005.
  • the batsman is dismissed, meaning the probability of winning drops by 0.05.
  • the overall batting match impact of this individual is:
  • the overall batting impact can then be multiplied by 100 to convert it to a percentage. This overall batting impact as a percentage is shown as "batting impact" in Table 4.
  • Table 4 shows an example of match impacts (from a T20 International Cricket match between England and New Zealand, 25 June 2013) :
  • a player has a positive impact if the outcome of the sporting event after the player comes onto the pitch or court improves.
  • the outcome of each interaction i.e. a delivery
  • the sign of the impact measure for the batsmen would be opposite to the sign of the impact measure for the bowler and fieldsman in a play. For example, if a batsmen in a play brings a -0.001 probability of winning for Team A, then the bowler in the same play brings a +0.001 probability of winning for Team B.
  • the engines and sub-models comprise computing devices, or combinations of computing devices, configured to execute specific sets of computer executable instructions.
  • the computing device is connected to other devices. Where the device is networked to other devices, the device is configured to operate in the capacity of a server or a client machine in a server-client network environment. Alternatively, the device can operate as a peer-to- peer or distributed network environment.
  • the device may also include any other machine capable of executing a set of instructions that specify actions to be taken by that machine. These instructions can be sequential or otherwise.
  • the term 'computing device' includes any collection of machines that individually or jointly execute a set or multiple sets of instructions to perform any one or more of the methods describes above.
  • the computing device includes a processor.
  • An example of a processer is a central processing unit.
  • the device further includes main system memory and static memory.
  • the processor, main memory, and static memory communicate with each other via a data bus.
  • Computing device may also include a reader unit, network interface device, display device, optical media drive, cursor control device, and signal generation device.
  • the reader unit is able to receive a machine readable medium on which is stored one or more sets of instructions and data structures, for example computer software.
  • the software uses one or more of the methods described above.
  • Reader unit includes a disc drive and/or a USB port.
  • the machine readable medium includes a floppy disc and a static storage device such as a thumb drive.
  • the optical media drive is used, the machine readable medium includes a CD-ROM.
  • Software may also reside completely or at least partially within the main system memory and/or within the processor during execution by the computing device.
  • the main memory and processor constitute machine-readable tangible storage media.
  • Software may further be transmitted or received over the network via a network interface device.
  • the data transfer uses any one of a number of well known transfer protocols.
  • One example is the hypertext transfer protocol.
  • the machine-readable medium may be a single medium or multiple media. Examples of multiple media include a centralised or distributed database and/or associated caches. These multiple media each store one or more sets of computer executable instructions.
  • the term 'machine-readable medium' includes any medium that is capable of storing, encoding or carrying a set of instructions for execution by the machine and that cause the machine to perform any one or more of the methods described above.
  • the machine- readable medium is also capable of storing, encoding or carrying data structures used by or associated with these sets of instructions.
  • the term 'machine readable medium' includes solid-state memories, non-transitory media, optical media, magnetic media, and carrier wave signals.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

La présente invention concerne un procédé pour déterminer un résultat d'un événement sportif. Le procédé comprend la détermination, par un moteur contextuel, d'un facteur de tendance à un quelconque instant de l'événement sportif, le facteur de tendance étant fondé sur des événements précédents dans l'événement sportif ; et la détermination, par un moteur de prédiction, du résultat de l'événement sportif an moins en fonction du facteur de tendance. L'invention concerne en outre un système conçu pour déterminer un résultat d'un événement sportif, et un procédé pour corréler une performance d'un joueur ou d'une équipe à une condition de jeu dans un événement sportif.
PCT/NZ2014/050012 2013-11-21 2014-11-21 Système et procédé pour estimer ou prédire un résultat d'un match dans un événement sportif WO2015076682A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
NZ617867 2013-11-21
NZ61786713 2013-11-21

Publications (1)

Publication Number Publication Date
WO2015076682A1 true WO2015076682A1 (fr) 2015-05-28

Family

ID=53179854

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/NZ2014/050012 WO2015076682A1 (fr) 2013-11-21 2014-11-21 Système et procédé pour estimer ou prédire un résultat d'un match dans un événement sportif

Country Status (1)

Country Link
WO (1) WO2015076682A1 (fr)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3608305A1 (fr) 2018-08-10 2020-02-12 Röhm GmbH Procédé de production d'acide méthacrylique ou d'esters d'acide méthacrylique
WO2020180702A1 (fr) * 2019-03-01 2020-09-10 Stats Llc Personnalisation de prédiction de performances à l'aide de données et d'une posture corporelle pour l'analyse de performances sportives
CN112766546A (zh) * 2021-01-05 2021-05-07 上海大学 一种基于机器学习的足球比赛胜负实时预测方法
CN113240190A (zh) * 2021-06-02 2021-08-10 郑州大学体育学院 基于多周期演化熵技术的运动员赛前状态评价方法
US11554292B2 (en) 2019-05-08 2023-01-17 Stats Llc System and method for content and style predictions in sports
US11577145B2 (en) 2018-01-21 2023-02-14 Stats Llc Method and system for interactive, interpretable, and improved match and player performance predictions in team sports
US11605268B2 (en) 2019-11-22 2023-03-14 Castle Hill Holding Llc System and method for wagering on past events
US11645546B2 (en) 2018-01-21 2023-05-09 Stats Llc System and method for predicting fine-grained adversarial multi-agent motion
US11682209B2 (en) 2020-10-01 2023-06-20 Stats Llc Prediction of NBA talent and quality from non-professional tracking data
US11918897B2 (en) 2021-04-27 2024-03-05 Stats Llc System and method for individual player and team simulation
US11935298B2 (en) 2020-06-05 2024-03-19 Stats Llc System and method for predicting formation in sports
US11948434B2 (en) 2021-02-02 2024-04-02 Castle Hill Holding Llc Method and system for conducting wagers

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8532798B2 (en) * 2011-08-23 2013-09-10 Longitude Llc Predicting outcomes of future sports events based on user-selected inputs

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8532798B2 (en) * 2011-08-23 2013-09-10 Longitude Llc Predicting outcomes of future sports events based on user-selected inputs

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
AKHTAR S. ET AL.: "Forecasting Test Cricket Match Outcomes In Play", INTERNATIONAL JOURNAL OF FORECASTING, vol. 28, 2012, pages 632 - 643 *
BAILEY M. ET AL.: "Predicting The Match Outcome In One Day International Cricket Matches, While The Games Is In Progress", JOURNAL OF SPORTS SCIENCE AND MEDICINE, vol. 5, 2006, pages 480 - 487 *
BOULIER B.L. ET AL.: "Predicting The Outcomes Of National Football League Games", INTERNATIONAL JOURNAL OF FORECASTING, vol. 19, 2003, pages 257 - 270 *
SCARF P. ET AL.: "Modelling Match Outcomes And Decision Support For Setting A Final Innings Target In Test Cricket", IMA JOURNAL OF MANAGEMENT MATHEMATICS, vol. 16, 2005, pages 161 - 178 *

Cited By (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11660521B2 (en) 2018-01-21 2023-05-30 Stats Llc Method and system for interactive, interpretable, and improved match and player performance predictions in team sports
US11645546B2 (en) 2018-01-21 2023-05-09 Stats Llc System and method for predicting fine-grained adversarial multi-agent motion
US11577145B2 (en) 2018-01-21 2023-02-14 Stats Llc Method and system for interactive, interpretable, and improved match and player performance predictions in team sports
EP3608305A1 (fr) 2018-08-10 2020-02-12 Röhm GmbH Procédé de production d'acide méthacrylique ou d'esters d'acide méthacrylique
WO2020030374A1 (fr) 2018-08-10 2020-02-13 Röhm Gmbh Procédé de production d'acide méthacrylique ou d'esters d'acide méthacrylique
WO2020180702A1 (fr) * 2019-03-01 2020-09-10 Stats Llc Personnalisation de prédiction de performances à l'aide de données et d'une posture corporelle pour l'analyse de performances sportives
US11679299B2 (en) 2019-03-01 2023-06-20 Stats Llc Personalizing prediction of performance using data and body-pose for analysis of sporting performance
CN113544697A (zh) * 2019-03-01 2021-10-22 斯塔特斯公司 用数据和身体姿态分析运动表现以对表现进行个性化预测
US11554292B2 (en) 2019-05-08 2023-01-17 Stats Llc System and method for content and style predictions in sports
US11610455B2 (en) 2019-11-22 2023-03-21 Castle Hill Holding Llc System and method for pari-mutuel gaming
US11605268B2 (en) 2019-11-22 2023-03-14 Castle Hill Holding Llc System and method for wagering on past events
US11651655B2 (en) 2019-11-22 2023-05-16 Castle Hill Holding Llc System and method for pari-mutuel gaming
US11710383B2 (en) 2019-11-22 2023-07-25 Castle Hill Holding Llc System and method for wagering on historical horse races
US11935298B2 (en) 2020-06-05 2024-03-19 Stats Llc System and method for predicting formation in sports
US11682209B2 (en) 2020-10-01 2023-06-20 Stats Llc Prediction of NBA talent and quality from non-professional tracking data
CN112766546B (zh) * 2021-01-05 2022-04-26 上海大学 一种基于机器学习的足球比赛胜负实时预测方法
CN112766546A (zh) * 2021-01-05 2021-05-07 上海大学 一种基于机器学习的足球比赛胜负实时预测方法
US11948434B2 (en) 2021-02-02 2024-04-02 Castle Hill Holding Llc Method and system for conducting wagers
US11918897B2 (en) 2021-04-27 2024-03-05 Stats Llc System and method for individual player and team simulation
CN113240190A (zh) * 2021-06-02 2021-08-10 郑州大学体育学院 基于多周期演化熵技术的运动员赛前状态评价方法

Similar Documents

Publication Publication Date Title
WO2015076682A1 (fr) Système et procédé pour estimer ou prédire un résultat d'un match dans un événement sportif
US11615676B2 (en) Method, system, and computer program product for interactive sports game
Stekler et al. Issues in sports forecasting
KR101585399B1 (ko) 스포츠데이터 정보를 이용한 경기결과 예측방법
Swartz Research directions in cricket
Scarf et al. The importance of a match in a tournament
Wolf et al. A football player rating system
Robberechts et al. Forecasting the FIFA World Cup–Combining result-and goal-based team ability parameters
Somaskandhan et al. Identifying the optimal set of attributes that impose high impact on the end results of a cricket match using machine learning
US11738269B2 (en) Interactive sports fan experience
Holmes et al. A Markov chain model for forecasting results of mixed martial arts contests
Soto-Valero et al. Assessing the causal impact of the 3-point per victory scoring system in the competitive balance of LaLiga
Zelenkov et al. Measuring the efficiency of Russian football premier league clubs
Munir et al. Predicting a T20 cricket match result while the match is in progress
McGarry et al. Winning squash: predicting championship performance from a priori observation
US20230245525A1 (en) Method, system, and computer program product for interactive sports game
Shoval et al. A probabilistic approach to the analysis of a volleyball set performance
US9192864B1 (en) Performance based baseball scoring system
Bouzarth et al. Storytelling with sports analytics
Manoj et al. American league baseball championship 2017 prediction using AHP
US6691063B1 (en) Measuring a baseball player's accumulated winning contribution
Hipp et al. Mining ice hockey: Continuous data flow analysis
US11872464B1 (en) Golf play outcome simulation modeling system
Team Forecasting the FIFA World Cup–Combining Result-and Goal-Based Team Ability Parameters
Spanias Professional tennis: Quantitative models and ranking algorithms

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14864565

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14864565

Country of ref document: EP

Kind code of ref document: A1