What games are the most crucial games in a series. At times a media outlet will trot out a statistic claiming that the team that wins game one wins the series at such and such a percentage. Is that true, and if so, does that make game one an absolute necessity. A bout of extreme boredom at work set me on the path of exploring some mathematical exploration. Using a statistical model of each team having a 50/50 shot of winning each game, I caculated how much the odds of winning a seven game series changed with each situation, such as a team up three games to one or two games to zero. Here are the various situations in order from most important to least. Importance here is defined by what extent the game changes the odds of winning the series.

1. Game Seven

2. Game Six

3. Game Four when one team holds a 2-1 series lead

4. Game Three when one team holds a 2-0 series lead

5. Game Five when one team holds a 3-1 series lead

6. Game Five when the series is tied

7. Game Two

8. Game Three when the series is tied

9. Game Four when one team holds a 3-0 series lead

10. Game One

In order to test this, I took a look at real life results, specifically MLB's last 45 postseason series. In the real world, teams taking 1-0, 2-0 and 2-1 series leads have won more often than we would expect. Teams taking a 3-0 lead have won less often, but that's entirely due to one outlier in 2004. Teams taking 3-1 and 3-2 leads have won almost <I>exactly</I> as often as we expect. I think the reason for the discrepancy in the first results is because when a team takes a lead, it is often because they truly are a better team, and thus are operating at odds better than 50/50.

Teams that have won game one since 1990 have won the series 71% of the time. That's significant, but then again teams that have taken game two as well have won 90% of the time, and that is another huge gain. I think if you looked at the results of any game, you would see the same result (except for game seven, which would give you 100%). You can shift the odds around to develop different percentages, but the order above would remain roughly the same.

I don't know if you can do anything with this, but I had it down on paper and figured I would print it.