PDA

View Full Version : For the stat heads...


Madscout
06-13-2008, 08:32 PM
Hey all. I've been playing around with some numbers, mainly standard deviation and how it pertains to baseball. I did a study a few days ago of all 30 teams, making a chart of ave runs and runs against, as well as standard deviation of runs scored and runs against and varience of runs scored and runs against. Let me explain what everything means.

I have the runs scored per game average and team ERA of everyteam listed, as calculated by me by taking all the scores of all 60 or so odd games played so far. I then calculated the standard deviation, which is to say, how much away from the average the team scores in a particular game. For example, a team that has played 6 games scoring the following runs: 7,3,7,3,7,3 has an average of 5 runs per game and a standard deviation of 2 runs.

I also have calculated the variance, which is commonly refered to as the "spread" of the numbers or what I call the "blow out/shut out factor". For example, a team playing six games scoring 7,3,7,3,7,3 has the same standard deviation as a team that has scored 7,3,8,2,6,4, but the second team should have a greater variance.

So far, I have found the study to simply explain in numbers how a team plays. For instance, the A's aren't really going to blow you out that much and aren't going to get blown out that much, but they will play a close game and pitch you well. They are going to bring a similar game every night. This is given by a relatively small standard deviation on both pitching and low variance in both.

Just wanted your thoughts. Thanks for reading...

Madscout

5826

SBSoxFan
06-13-2008, 09:53 PM
I didn't read the doc, but I agree with your idea, and mentioned it on here a couple years ago. This is one of the many reasons the Pythagorean win stat is rather useless --- it doesn't account for standard deviation in runs scored.

:thumbsup:

jabrch
06-14-2008, 01:47 AM
Looks like if you compare to teams in the same quartile as us in terms of RS we are well above the mean in terms of SD - but only by a few tenths - which is fairly small.

Thanks for the work - I was hoping someone would do what I was too lazy to do myself.

ondafarm
06-14-2008, 09:51 AM
With a superior pitching staff like ours, there is no reason why we should not have a lower SD for runs against. I believe this is a clear indictment of Ozzie Guillen's management of the pitching staff.

As for the runs scored, the larger the variance, the more explosive the offense, but also the more it depends on the long ball. Good teams know how to manufacture a run or two every game. Again, if this team was committed to small ball as well as the long ball, we would have a better record and lower variance.

munchman33
06-14-2008, 10:36 AM
Looks like if you compare to teams in the same quartile as us in terms of RS we are well above the mean in terms of SD - but only by a few tenths - which is fairly small.

Thanks for the work - I was hoping someone would do what I was too lazy to do myself.

We're talking about standard deviation over a large sample size. Tenths of a point are huge.

FWIW, look at the Angels offense.

gosox41
06-14-2008, 10:21 PM
Again, if this team was committed to small ball as well as the long ball, we would have a better record and lower variance.

Would that be nice. Too bad we're not committed.


Bob

Madscout
06-15-2008, 01:02 AM
With a superior pitching staff like ours, there is no reason why we should not have a lower SD for runs against. I believe this is a clear indictment of Ozzie Guillen's management of the pitching staff.

As for the runs scored, the larger the variance, the more explosive the offense, but also the more it depends on the long ball. Good teams know how to manufacture a run or two every game. Again, if this team was committed to small ball as well as the long ball, we would have a better record and lower variance.

I agree with you that we would like to have a lower standard deviation, but look at it in combination with the variance. The high variance would seem to say, that when we pitch poorly, then we get blown out, giving up 10+ runs or so. It says also that about the same that we give up 10+ runs, we shut out or only give up 1 run. Or it says that we pitch all along the spectrum, however out standard deviation is better than many good teams, including the angels. You have to think of standard deviation and variance as both above and below the mean or average.

I also agree with you that the angels' numbers offensively would be nice. They seem to get their 4 runs most nights, and with the way they pitch, they are a tough team to beat.

I also have to say that it is interesting to see how nice the A's numbers are, which confirms to me that they are really looking to maximize such things.

All in all, the average is still the most important number. I am finding that when I start looking at variance and standard deviation, I am always forced to quantify it through the average. Really my thought process goes like this, low/high average-> good/bad -> standard deviation-> consitant or not-> variance-> does this team have consistant nights or extremely good/bad nights.

jabrch
06-15-2008, 01:04 AM
Again, if this team was committed to small ball as well as the long ball, we would have a better record and lower variance.


But we'd lose some of the games that we are currently winning also....

It is a give and take dynamic.