Baseball Totals

[tech]

[ QUOTE ]
I am surprised by the responses I have gotten. I am trying to take a mathematical approach to computing baseball totals. I have yet to see anyone make a single argument against my approach.


[/ QUOTE ]

It's nothing personal. We are trying to help save you hundreds of hours of work by telling you the things we learned the hard way.

I don't think you fully appreciate the value of nonparametric and/or distribution-free statistics. You are basically assuming a symmetric distribution of runs in all your calculations. That simply is not the case, and it is not close enough to assume the difference is negligible. You should be trying to calculate the probability that a game will go over the posted total. To do that, you need to approximate the distribution of the runs scored. If your assumption about the distribution is wrong, it doesn't matter if your expected runs calculation is perfect in every other way.

[Paluka]

Your post seems really obvious to me. It would be a pretty stupid assumption to think that just because something on average is 9.5 that the results are necessarily distributed 50% below and 50% above.

[sportypicks.com]

Doug,

I'm glad that you post topics like this as I prefer reading topics on methodology rather than the run of the mill stuff you get on most message boards (Play this, fade this, etc.).

As far as your measuring ballpark affect goes, I believe you can get data like this off of statistical sites out there. Otherwise you can always create it yourself depending on the purpose you are using it for. Ballparks are something that don't change from year-to-year like teams do, so you can use a few years of data to calculate an accurate ballpark factor. I wish you well in your simulations, and I'm interested in hearing how you deal with bullpens.

Bullpens is what took the longest amount of time in our handicapping methods. I don't believe there is a "right" answer and I'm sure there are many ways to deal with it.

Good stuff.

[DougOzzzz]

[ QUOTE ]
[ QUOTE ]
I am surprised by the responses I have gotten. I am trying to take a mathematical approach to computing baseball totals. I have yet to see anyone make a single argument against my approach.


[/ QUOTE ]

It's nothing personal. We are trying to help save you hundreds of hours of work by telling you the things we learned the hard way.

I don't think you fully appreciate the value of nonparametric and/or distribution-free statistics. You are basically assuming a symmetric distribution of runs in all your calculations. That simply is not the case, and it is not close enough to assume the difference is negligible. You should be trying to calculate the probability that a game will go over the posted total. To do that, you need to approximate the distribution of the runs scored. If your assumption about the distribution is wrong, it doesn't matter if your expected runs calculation is perfect in every other way.

[/ QUOTE ]

The assumption about distribution is not wrong, however - it appears everything I've done has been done before. Go to www.tangotiger.net and scroll down to the bottom. The "tango distribution" at the end does pretty much the same thing as my excel spreadsheet. The results are very close to real life numbers. Read the discussion there if you want more info. My spreadsheet uses the same formula.

The whole point of the exercise was to take "averages" and convert them into actual run distributions.

[DougOzzzz]

[ QUOTE ]
Your post seems really obvious to me. It would be a pretty stupid assumption to think that just because something on average is 9.5 that the results are necessarily distributed 50% below and 50% above.

[/ QUOTE ]

Yeah, but I thought it would be interesting and useful to see what the actual distributions were, based on expected RPG.

[DougOzzzz]

[ QUOTE ]
Doug,

I think you misunderstood me. I didn't say YOU were lazy, I was talking about people who just pick up the paper and read batting averages and era's and then bet on a game.

I probably should have conveyed myself better. I think you are definitely on the right track as far as producing accurate total numbers. All of those factors you mentioned above do come into play when handicapping totals.

My question for you is...How do you measure pitcher's strength? By just looking at his ERA (pitching avg) that would be the wrong way to go. One bad outing where he gets pulled after 1/3 of an inning giving up 5 runs can really distort an ERA. That is what Tech was talking about. Instead of taking an average, compile a distribution based on number of innings and runs allowed during these innings. I think you will fare much better in your simulations.

[/ QUOTE ]

I'm not too crazy about using a pitcher's ERA in the first place - K/BF, BB/BF, and HR/BF correlate much better to future ERA than ERA itself does.

As far as I can tell from the distribution thing you are trying to say that analyzing a pitcher's OWN runs allowed/inning distribution is more useful than taking the average Runs per inning distribution of pitcher's expected to allow X runs/game. I would say that the sample size would hurt this significantly. I'm sure two equal pitchers can probably have different run distributions, but that type of thing would take hundreds of thousands of innings to really show. And I am sure there are "reasons" that pitcher's would have run distributions that don't correspond to this model. An extreme groundball/control pitcher who doesn't strike out alot of guys may have the same ERA as a hard-throwing guy who gives up his fair share of walks and home runs, but the first guy is probably going to allow 1 run to score more often and avoid big innings better than the 2nd.