Two of the most valuable tools for the Fantasy General. Here are the basics.

I have received a few inquiries on out use of hit rate and strand rate and why we use them in every player analysis, so I wanted to address these two stats.

It all started with Voros McCracken. In one of the most famous and far-reaching discoveries in baseball, and certainly one of the most important in statistical analysis, he discovered that pitchers do not exert control over the actual number of hits they allow.

With his discovery, which he called “defense independent pitching,” McCracken essentially found that the percentage of balls put into play against a particular pitcher that fell for hits did not correlate well across seasons. This implies that controlling the number of hits is beyond the pitcher’s control. This discovery has been verified, and slightly altered in that pitchers may have some slight influence, but the overall effect is one of randomness.

Fantasy baseball, at least for those who heed this finding, has never been the same. his discovery means that pitcher’s hit rates will tend to regress, assuming all else to be equal, to a mean, generally 30%, or a .300 batting average on balls in play. So what does one do with this information?

Let’s examine it by analogy. You are watching a blackjack tournament. You are tracking the dealers and charting how many blackjacks each dealer deals. You find that certain dealers, lets say 5% of them, are dealing far more blackjacks than the rest. What is your conclusion about these dealers?

I hope it isn’t that they are skilled at dealing blackjacks, or that they are cheating!! In any population that is performing a task that is randomly distributed you will get a certain number that are outliers. This is measured by standard deviations and the bell curve. You would expect these dealers to regress to the normal mean of dealing blackjacks.

A pitcher is nothing more than a blackjack dealer, at least when it comes to the number of hits on balls in play. The pitcher CAN control how many balls are actually put into play. But once it is in play it is left to the fates. It helps if the pitcher’s fates have a good defense and high grass!

Note that this does not mean that every pitcher will regress to the 30% mean. It means that all else being equal they will regress to the mean. The example of Scott Kazmir from my earlier article shows this principle. Defense, parks, and other factors may intervene, since all else may often not be equal. This is the Fantasy General’s task; to find out.

The value of a hit rate is that it allows the Fantasy General to make a determination of whether a pitcher is lucky or unlucky. This is the heart of fantasy baseball; finding out who is underrated, and more importantly undervalued. I may have beaten this to death on the blog so far, since I address these factors in every player evaluation, but the underlying principle permits us to understand its power.

When we address luck we don’t necessarily mean just whether a bloop falls in or not. It also refers to a player outperforming what his skills say his performance level should be.

What about the strand rate? Well, if we know that hits are randomly distributed we know that strand rates should vary randomly also, at least for the most part. Granted there may be/are pitchers who are able to bear down and lower their strand rate because they may allow fewer balls in play with runners in scoring position. So it is not all luck, and there is some wiggle room. But not a lot.

Why? Even these skilled pitchers can only do so much. The distribution of hits, namely when and in what situations they occur has a great impact on the pitcher’s ERA. This distribution will be influenced by the number of hits when a ball is put into play, which is essentially random per McCracken.

The best pitchers will have strand rates approaching 75%. The worst will be in the range of 65%. Anything above or below is a strong candidate for regression. Relievers of course can vary wildly since their innings pitched are much smaller and subject to the vagaries of chance more significantly. But the overall range from the best to the worst can be somewhat limited especially among starters.

For example, currently Dan Haren has the best strand rate among starters at 81%. The worst is Horacio Ramirez at 60%. Strand rate is a key indicator for the quality of a pitcher. It also will tell you whether a pitcher is lucky or not in relation to his particular skill set. That is the key to value and the bedrock of fantasy baseball.

How will they regress and of what value is this to the Fantasy Player? One comment I received recently pointed out, correctly, that luck can even out over a large sample size in a smooth fashion, without large changes in the overall performance picture.

This is theoretically true, and is certainly true if you are a poker player for example and will play millions of hands. But in the short term the regression/correction is typically rough. The “short term” includes the remainder of a given season; one season is still a relatively small sample size. Because of the inherent variability and statistical noise the regression is usually rough, and is not smooth within a given season. Even a cursory glance at the vast majority of players shows that the correction is generally rough and not smooth.

This is what the Fantasy General is gambling on when buying low and selling high, or trading a “lemon,” a correction that will occur that is significant enough to tilt the balance in his or her favor.

When Pitcher X has a hit rate of 25%, the astute Fantasy General looks to trade him. Why? Because we expect regression. That is not to say that he will have a 35% hit rate to balance out to 30%, though this may happen and often will. But it is also grounded in the fact that the typical owner is not aware of this information, and will value the player based on his 25% hit rate. If so, then you gain since you get 25% hit rate value for a player who will have a higher hit rate going forward. It may be 28% it may be 30% or 40%. Marginal value is still obtained.

By itself these stats will not lead to victory. As Sun-Tzu stated, if one who finds the majority of factors are in his favor will be victorious, and one who finds few factors in his favor will be defeated, what of one who finds no factors in his favor? Finding the multitude of factors in your favor is where the skill and judgment come in to play and where the best owners will shine.

My thanks to Ron Shandler and Phil Birnbaum for their input in crystallizing my thoughts for this post.

articleId=’15643′;