The Wicky T20 Match Simulator is an interactive tool that aims to simulate or predict viable results of upcoming T20 games. It predicts individual runs scored, individual bowling, team runs scored/conceded, and boundaries scored by a team and individual.
In the current sports analytics landscape, it is a tool with great potential for T20 franchises, broadcasters, and spectators. You can find our T20 Match Simulator here.
Through our market research, we’ve noted that although analysts are aware of data science and machine learning algorithms, most teams don’t have the time or budget to go beyond traditional historical analysis. Predictive analytics and machine learning is just a pipe dream for most franchises. This is something we hope to change soon.
How does our T20 Match Simulator work?
As mentioned earlier, we are looking to make feasible predictions for T20 games. This presents quite a challenge. Apart from actually predicting the game, we are trying to simulate it multiple times and find an aggregation. This is so that it gives us a range of numbers in which the eventual number could fall.
Without going into much technical jargon, the way we currently simulate a game is by inputting historical data to a transition matrix per ball of a T20 game. The transition matrix covers the possible ‘states’ in which the ball could be i.e. 0, 1, 2, 3, 4, 6, wicket, extra. Aggregating these states forms innings totals and individual statistics.
From our research, we find that 100 simulations are optimal. This maintains a good balance between variation in the results and avoiding results that tend to averages.
One of the challenges we have faced is the balance between accuracy and practicality. Realistically, we need to run our simulations in the 5-10 minutes after lineups are announced at the toss. This is possible for 100 simulations currently. However, if we look to run 1000 simulations, the time constraint becomes more of an issue.
Check out our T20 Match Simulator for the Indian Premier League (IPL)
What are the possible use cases?
T20 Franchises
The T20 Match Simulator could be vital for T20 franchises in the current landscape with an increased emphasis on data. It could enable teams to assess par scores at a certain venue. Ahead of a game, this information could help strategise batting approaches to surpass these marks. Teams will be able to see how different batting/bowling combinations will fare in an upcoming game according to historical data.
The T20 Match Simulator can be particularly useful in edge case selection scenarios. These include choosing between a spinner or a pace bowler at a particular venue or against a particular opponent. It can provide insights as to what may happen if different lineups are played.
Broadcasters
Broadcasters are increasingly showing a wider range of data and charts during their coverage. A majority of this pertains to historical data of players and venues. This is undoubtedly a reasonable way to assess an upcoming game. However, there is nothing as yet in the industry that looks to predict an upcoming game.
Although win probability models like WinViz dynamically change as the match progresses, we think a prediction of what is likely to happen before a game is innately more interesting to decision makers and fans alike.
Fantasy Sports & Betting
Fantasy sports companies and their users could use the T20 Match Simulator to make better decisions for selecting their fantasy team. A simple way they could do this would be to see if the Match Simulator shows a team with a unusually high likelihood of scoring 180+ runs or a player scoring 50+ runs. Considering this takes in historical player, team, and venue stats, it would help users make decisions on their fantasy teams or pre-game betting.
Furthermore, fantasy players must often lock in their teams before the start of the match. They cannot make any changes to their team when the match is in progress. Thus, the insights from the T20 Match Simulator are especially valuable in such a scenario.
T20 cricket is dynamic and unpredictable. However, with probabilities and historical data, we are able to offer immense value with this tool to various parties.
We’ve also got a version of our T20 Simulator for the Pakistan Super League (PSL)
What else is available in the market?
Historical data have long been available to the general public, with providers like espncricinfo.com and the Statsguru tool, live since 2000. T20 franchises around the world, in leagues such as the BBL and the Indian Premier League (IPL), have used this data in combination with improvements in sports data science to gain a competitive advantage over opponents.
Franchises are increasingly open to using data to gain a competitive advantage. Through our market research, we’ve noted that although analysts are aware of data science and machine learning algorithms, most teams don’t have the time or budget to go beyond traditional historical analysis. Predictive analytics and machine learning is just a pipe dream for most franchises. This is something we hope to change soon.
Predictive tools such as WinViz and Winning and Score Predictor (WASP) are available to broadcasters and teams alike to make decisions as the game progresses. However, several models, including the two above, change predictions dynamically based on game state. From our research, nothing in the market is as yet available to make viable predictions for upcoming games based on historical indicators and aggregations of multiple simulations.
If you are interesting in getting to know more about the T20 Match Simulator or any of our other work at Wicky, please reach out via email or on social media.