This will be a quick-ish post on the basics of game theory. However, rather than opting to use the standard ‘Prisoners’ Dilemma’ scenario, I felt a more topical situation would be fitting and will hopefully make for an interesting application of game theory in the real world. Let’s also assume that the vote hasn’t already happened, too.
In a scenario where each candidate can vote for as many other candidates as he wants, the aim of the game for myself is clear: I need to accumulate more votes, whilst it is also preferable if the other candidates score as few votes as possible in comparison. Subsequently, the temptation here is to create ‘vote exchanges’ where two candidates each agree to vote for each other. However, game theory can help explain why the candidates will often (or every time in a purely hypothetical world) not stick to their word.
The first step is to determine ‘payoffs’ for the different outcomes, measuring the utility that the candidate receives following each outcome. In this case, the players shall be Person X and Person Y (unimaginative, I know). The best outcome for X is for him to not vote for Y, but for Y to vote for him, and thus he has an extra vote and is more likely to triumph overall. The next-best outcome is for both applicants to vote for each other. Although the vote is diluted somewhat, it is obviously better for X to have a vote than to not have a vote. Then, it is better if neither candidate votes for each other – this means that while X gains no advantage overall, he at least does not have to face Y having an advantage with respect to the rest of the competition, which occurs in the worst-possible scenario for Person X. Here, he gains no votes whereas, as a result of voting for Y, Y gains a vote.
The utility will work in the exactly the opposite way for Person Y. We can consequently number the outcomes, with 3 being the most preferable outcome and 0 being the worst. The payoff matrix looks like this:
| Vote | Not Vote | |
|---|---|---|
| Vote | 2,2 | 0,3 |
| Not Vote | 3,0 | 1,1 |
Note that Person X’s choices are on the left, while Y’s are at the top. X’s payoff utility is the first number in each box.
Now, if we look at the table the (Vote, Vote) payoff is clearly the most ‘fair’ and ‘economical’, since the combined payoff is the highest of all the outcomes. However, analysing each person’s individual choices can help understand why the (Vote, Vote) outcome will not occur. Assuming X knows for sure that Y will vote for him, here are his options:
| Vote | |
|---|---|
| Vote | 2,? |
| Not Vote | 3,? |
Knowing that Y will vote, X can either vote or not vote. If he votes, his payoff will be just 2, whereas by not voting he can achieve the maximum payoff of 3. Therefore, he will chose ‘Not Vote’ when he knows that Y will vote. Here’s what will also happen if he knows that Y will not vote:
| Not Vote | |
|---|---|
| Vote | 0,? |
| Not Vote | 1,? |
X voting will provide the worst payoff, but he can achieve a better payoff of 1 by voting. As he will, again, choose to not vote, we can call ‘Not Vote’ the strictly dominating strategy for Person X, because it is better for him to not vote, regardless of Person Y’s strategy. The same is true for Person Y – if he knows that X will vote, he will similarly achieve a better payoff by not voting, and if he knows that X will not vote, he will once more achieve a better payoff by also deciding to not vote. ‘Not Vote’ is subsequently his strictly dominating strategy, and as a result the players arrive at the (Not Vote, Not Vote) strategy. This is known as the voting system’s Nash Equilibrium, because neither player has any incentive to deviate from their current strategy. The ‘vote-for-vote’ deal is rendered a waste of time.
Of course, in reality there are other factors to consider. For example, if the candidates are both good friends then they will feel more benefits in voting than not doing so, meaning actual payoffs are different. However, it is interesting to see how game theory can be modelled to explain very relevant, real life phenomena – and, for anyone out there having to vote for other people for a position, never trust someone who promises to vote for you!






