Sunday, 5 October 2014

October Strategy

The last couple of nights / mornings have been good for those opposing the favourites in the MLB play-offs, and I am indebted to Fizzer555 for pointing me in the direction of some historical stats to validate my thoughts posted this week.

The recency effect is well known in betting, but I have written before that in baseball, the shortest of favourites in baseball is priced longer than you would find in other major sport, because on any day, any one team can beat another.

As mentioned in this excellent article on play-off baseball by Jeff Fogle:

Many in the field of analytics emphasize the role of randomness in the baseball post-season, emphasizing that “anything can happen in a small sample size” and “the best team usually doesn’t win the pennant.”
The best team certainly doesn't. The best team (the one with most wins anyway) this year in the National League was the Washington Nationals. They trail 0-2 in a best-of-series and now travel to San Francisco needing to win both games there to stay alive.

The best (most wins) team in the American League was the Los Angeles Angels of Anaheim - they are now in Kansas City trailing 0-2, fighting for their lives, and 'dogs in today's game.

So understanding that sample sizes are small, how has backing the underdog in recent seasons performed as a system?

Underdogs can be home teams or away (road) teams of course, as there are no games at neutral venues in the MLB, and there are more road 'dogs than home 'dogs. Going back to the 2011 season, the numbers for all 'dogs are:
While a couple of years have shown losses overall, both were small, and backing home 'dogs has been a winning strategy in all three years. Fizzer went back even further than me, and found over 10 years an ROI on all 'dogs of 4.8%, with home 'dogs at 13.2%. Since 2011, the numbers show each game is essentially a coin-flip.

It's simplicity reminds me of the Bundeslayga system, but whereas that system takes advantage of favourites driven too low in a league where strange things happen, and have happened over many seasons, the MLB Play-Off 'dog system takes advantage of favourites driven too low, perhaps by less savvy bettors anxious to enjoy the drama of the play-offs with a financial interest.

Thanks again to Fizzer Tony for his much appreciated input.

Update: The screenshot includes the two results from last night. Hot favourites Washington Nationals were ~1.5 pre-game, traded at 1.01 with a one run lead, and lost to the road 'dogs San Francisco Giants in a marathon game of 18 innings - the longest play-off game in MLB history by time, and tied by number of innings, and MLB has a long history.

In the other game, the Los Angeles Dodgers were pre-game favourites at ~1.61, traded at 1.07 with a two run lead, gave up the two runs in the top of the 8th, but the underdog St Louis Cardinals were finished off with a home run in the bottom of the 8th. No underdog winner on this occasion, but close with the Dodgers trading as high as 2.0 in that 8th inning.

No comments: