Is their a totals bias after an extra inning mlb game?

[ANDTHEWINNERIS]

DO TEAMS OFF AN EXTRA INNING GAME TEND TO PLAY UNDER IN THEIR NEXT
GAME?:



A subscriber e-mailed me, and asked me about this, as he said that most
recently at least, the answer appeared to be yes. I finally had a chance to dig
in a little deeper and found the answer, which will be the topic of today's
newsletter.



Would it make sense that perhaps the next day's starting pitcher is perhaps
the only rested player in a game that was obviously tight, required some extra
focus, and left all the players a bit emotionally, as well as physically
drained, going into play the next day?



The premise on a stand alone is a vig eater as the under record stands
at:



1033-918-87 for a win rate of 52.9% which shows profit as a stand
alone.



Digging a bit deeper when the home team is off a loss after an extra inning
game the results take a quantum leap:



433-320-34 for a win rate of 57.5%!! Considering the fact that this is a 10
year sample of nearly 800 games, that is a pretty strong situation. This has an
effective z-score of 4.03, which indicates a situation that is highly
predictive, and will likely continue to provide value.



This situation is enhanced if that extra inning game played under, as it
goes to:



182-127-18 for 58.9% winners.



BUT BUT BUT:



despite what looks to be a better system cashing 58.9% the z-score actually
drops to 3.04 which is still strong, but in this case playing the original vs
the subset is a much more profitable way to go!

[Jefferey13]

Interesting, good post.

[SlickRick1382]

Bk'ed for future reference / research ...

[whtsox13]

Next game after extra inning game, you are looking at 'team total' or 'game total' under?

[ANDTHEWINNERIS]

game total

[whtsox13]

game total

So the assumption is the extra inning game isn't the last game of the series and that the same teams will play each other the next day? TT under may be worth a look depending on where in the series the 'long' game falls.

[husky]

So the assumption is the extra inning game isn't the last game of the series and that the same teams will play each other the next day? TT under may be worth a look depending on where in the series the 'long' game falls.

This is a good point whtsox. Same matchup or different matchup?

[cpboy99]

So does this mean Az should be under tmrw? I like Corbin already but the Phils pitcher pitched for extra innings so not sure who will step in.

[ANDTHEWINNERIS]

Don't try to pick and choose and read more into what is there, this becomes back fitting variables that are inconsequential to results, the fewer back fits, and the better the sanity of the premise, and the greater the sample size, the better the results maintain themselves in the future. these are guidelines to offer an indication of where the value is likely to be.

[ANDTHEWINNERIS]

This is how people get in trouble, they look for subsets that are better, and end up doing worse. This as stated has a 4.03 z-score (and a better F-score which is even better for predictability using degrees of freedom, but that is way too complicated to explain here). There is a subset of this situation that is 65-28-5 which is 70%, but here is the problem. That has a z-score of 3.74, which has fewer games, less profit potential, and less certainty of performing at the same rate in the future. My advise, backed by statistical analysis, is don't pick and choose because it looks better, because it isn't!

[ANDTHEWINNERIS]

z-score measures DPMO (defects per a million opportunities, or a statistical confidence going forward of a situation. For example a DPMO of 0.5 means 308,000 defects in a million opportunities, or a not so certain <70% confidence. A DPMO of 2.5 has just 6,210 defects per a million, or 93.79% confidence, I like >95% to 98%.