MLB Totals Thread

[barman]

BASEBALL, like all major league sports is extremely actuarial. The Deep Stats guys who have emerged during the past 10-15 years are replete with rational analysis which demonstrate how all stats will fall into a reasonably predictable range and that many categories of stats can be correlated with regard to their expected results over a sufficiently large sample.

 

MLB as a whole (season of 2460 games) averages 9 runs per game with a seasonal variance of +/- .3 runs

 

General actuarial results can be reasonably forecasted with a sample as small as 600-800 events with strongly similar ground rules. In 800 games for example, the number of "high scoring" teams are counterbalanced by "low scoring"....Home and Away events deliver inverse balancing, injuries to either pitchers or batters effectively counter each other out.

 

--------

Nothing written above is revolutionary. But how can we use that foundation to give us a bit of help when forecasting results of upcoming events?

 

Well if "9" is the MLB average and a sufficient percentage of games go high or low to that baseline we can reasonably conclude there will be a regression back towards the "9" the more games we add to our summary

 

This just completed weekend of 45 MLB games saw just nine games tally 10+ runs. Five games tallied exactly 9 and 31 games delivered less than 9.

 

Par for 45 games is 405 runs. This weekend only 334 runs scored.

[KingRevolver]

Time to start pounding Overs?

[Miley Cyrus]

Ya know the record to date this year or records of previous years?  Too lazy to look it up.  Regression in theory should occur assuming the oddsmakers adjust the totals low enough based on past results.  Unlike the logic of people expecting to see more results of "heads" on upcoming coin flips because the last million in a row have been tails.