Undergrad Battle Projections: Predicting Probabilities in Games

[Kaura]

This is a rather odd topic but recently when playing games, mostly first person shooters, I have formed a curiosity about "Battle Projections" or the ability to predict probabilities based on in game variables.

For example, if you were spectating a round of no respawn four versus four death match, would it be possible to use compiled data such as each player's kills and deaths and the remaining alive players to calculate the probability of each team winning?

I know they do this in certain sports, so I suppose my main question is how to go about this and what exactly I should study to learn more about it.

 

[fresh_42]

I know they do this in certain sports, so I suppose my main question is how to go about this and what exactly I should study to learn more about it.

The obvious answer is stochastic aka probability theory and the hardest part might be to find out the given settings by the game itself, i.e. the phase space in which your calculations will take place. But there is more to it which points to decision and game theory: the personal risk aversion functions of the participants. Two players in exactly the same situation might still get to a different result evaluating this situation and therefore will chose different strategies. Overall I think there is a vast of unknown variables which you will have to make assumptions on before you could even start to make predictions, resp. projections and calculations.

 

[Kaura]

Cool I will look into it.

 

[jedishrfu]

There was a TV show called Deadliest Warrior which used software to predict the outcome of a match between two historical warriors like Spartan vs Ninja, Shaolin Monk vs Moari Warrior.

https://en.wikipedia.org/wiki/Deadliest_Warrior

and more about the sim used:

http://www.martialdevelopment.com/blog/deadliest-warrior-combat-simulator/

Many people agreed it was primarily for entertainment and not very realistic but perhaps you can learn something from it.

A Pipeworks sim too:

https://en.wikipedia.org/wiki/Deadliest_Warrior:_The_Game

 

[scottdave]

I just stumbled upon this question, although it is not new. You may find the book Math Bytes by Tim Chartier, interesting. One chapter discusses math techniques which can be used to evaluate strengths of competitors.