Thursday, September 05, 2013

How Important Is Playoff Home Field Advantage For The Dodgers


A post from Mike Scoscia's Tragic Illness titled 'Does The Best Record In The League Matter' got me thinking about a good way to measure this using math. Hey, opinions are great but if you know a little bit about me, you would know that I need to know the probabilities.

So what I did was assign a string of probabilities for the Dodgers winning the NLDS, NLCS and World Series. Now, I am going to assume in this exercise that the Dodgers are either the #1 or #2 seed in the National League. It is of course very possible that the Dodgers get the #3 seed, but let's save that scenario for a later date.

I kept the probabilities of the Dodgers winning the NLDS and World Series constant and then set a game by game win probability for each of the seven games the Dodgers and Braves would play each other in the NLCS. One set of probabilities for the Dodgers having HFA and one set for the Braves having HFA. The accuracy of the probabilities that I set aren't all that important. The important part is adjusting the NLCS games by 9% (assuming a HFA of 4.5% in this exercise) on a game by game basis. From these seven NLCS win probabilities I can then run this through a simple Monte Carlo simulation to get a NLCS series win probability.

The table below will show you my findings.

LAN HFAATL HFA
NLDS0.640.64
NLCS G10.650.56
NLCS G20.620.53
NLCS G30.470.56
NLCS G40.450.54
NLCS G50.560.65
NLCS G60.620.53
NLCS G70.560.47
NLCS Winner0.6350.607
W.S.0.50.5
W.S. Champion20.32%19.43%

For the data above you can see that given the series and game win probabilities that I used, there is less than a one percent difference in chances of winning the World Series based on having the #1 or #2 seed in the MLB playoffs. I think the numbers I used are pretty accurate though, as Vegas has the Dodgers chances of winning the World Series at between 19-20 percent. Anyways, the final number doesn't really matter. What matters is the difference between the chances of winning the World Series with the #1 or #2 seed. Is a cost of 0.89% to your World Series champion probability a big deal or not? I would say no it isn't when you are balancing that out with properly getting your team ready for the post-season.

Note: NLCS Winner was computed with Monte Carlo simulation (10 Million simulations).

2 comments:

Anonymous said...

The problem with this analysis is that it assumes a constant 64 percent chance of winning the NLDS, regardless of which team is the top seed. It should be clear that the chances of winning the NLDS are significantly greater if you are the top seed playing the wild card team, which played two days earlier and probably burned their top pitcher to win the wild card game, than if you are the second seed playing another well-rested division winner who was able to set up their playoff rotation the same way you do.

Xeifrank said...

Anon, that is a fair point and thanks for the thoughtful comment. But there are a lot of assumptions you could make. You just don't know how the WC teams rotation at the end will shake out. Plus they have a day off before the WC game and a day off afterwards.

So I took a situation where the WC used up their #2 starter in the WC game and reran the numbers. Instead of a 1v1, 2v2, 3v3, 4v4 starting pitcher matchup for the 7 game series, I used a 1v3, 2v1, 3v4, 4v2 and it bumped up the Dodgers chances of winning the NLDS by 3.63%. So I changed their 64% chance of winning the NLDS to 67.63% and it raised their chances of winning the World Series to 21.47%.

So you have the...
19.43% with ATL having HFA
20.32% previous LAD having HFA
21.47% new LAD having HFA

So, just over a 2% increase in odds. You can probably take worse case scenario and increase this by a fraction and you can take best case and drop it by a little bit. So I'd take a swag that the range is from 1.5% to 2.5% increase in chances of winning the World Series by having the #1 seed over the #2 seed.