A team of computer scientists has solved a “Colonel Blotto” game theory scenario which can be used to predict the outcomes of two-party conflicts.
The game was invented in 1921, but received its “Colonel Blotto” tag after WWII when scholars in operations research applied it to battlefields.
In their version, Colonel Blotto and his opponent are tasked with distributing a certain numbers of soldiers over a certain number of battlefields.
The rules are simple – whoever allocates the most numbers of soldiers to a battlefield wins; you don’t know how many soldiers your opponent has allocated; and the winner is the player who wins the most battlefields.
It’s considered a classic game theory scenario because solving it can unlock a powerful tool for making informed choices in areas such as political and military strategy, investment decisions and even in developing a rival consumer product.
And while variations of the game have been solved, until now no one has been able to find a way to arrive at a definitive solution for a two-party scenario.
Now, a team from the University of Maryland, Stanford University and Microsoft Research says it has developed “a generalised algorithm, which can now be applied to specific scenarios, such as the 2016 presidential election”.
“Our algorithm can potentially be used to compute the best resource investment strategy for any competitor up against a single opponent,” project lead Mohammad Hajiaghayi, associate professor of Computer Science at UMD told PhysOrg.
“As long as we have sufficient data on a given scenario, we can use our algorithm to find the best strategy for a wide variety of leaders, such as political candidates, sports teams, companies and military leaders.”
Hajiaghayi said the algorithm works as a proof-of-concept which the team plan to test in the US election later this year.
It will report its full findings at the Association for the Advancement of Artificial Intelligence (AAAI) Conference in Phoenix on February 15, 2016.