Log in Register

Login to your account

Username *
Password *
Remember Me

Create an account

Fields marked with an asterisk (*) are required.
Name *
Username *
Password *
Verify password *
Email *
Verify email *

fb mb tw mb

Wednesday 22nd Feb 2017

Perhaps the most asked question at this point of the season is exactly how long should one wait for a player to perform as expected?  How long should one wait before cutting bait?  Conversely, how long does a player have to exhibit a better than expected performance before it can be considered real?  While this discussion is not going to broach this exact subject, it is going to be in the neighborhood.  But for those interested in reading a rather involved study on the topic, I invite you to check out THIS PIECE, originally published by Pizza Cutter in 2007, recently posted again at www.fangraphs.com.  In it, the author determined the number of plate appearances necessary for a bevy of metrics to stabilize, that is be considered real, for better or worse.

What I want to talk about today is related and is more along the lines of what is commonly referred to as The Gambler’s Fallacy.  In layman’s terms, say you flip a coin twice and both times it comes up heads.  You are then asked to predict what will happen if you flip it two more times.  If you suggest it will land on tails both times, balancing the heads and tails for the four flips, that is an example of Gambler’s Fallacy.  In terms of chance, what is done is done and has no bearing on the chance going forward, so the proper response is to expect one head and one tails.  This is not to say it will definitely happen, but the probability of it happening is the greatest of all the possibilities.  In short, luck does not even out.

What does this have to do with fantasy baseball?  I feel more comfortable proposing the following as a hypothesis as opposed to a fact, as I have not yet “done the math” to prove it.  My contention is that when it comes to player performance, over the course of the long season, a player’s skill will find its proper level, but while a luck-based stat should regress, the luck, good or bad, will not flip.

Let us first look at a practical example to make sure everyone is on the same page with this hypothesis.  Without going into the details just yet, if you play the game at an intermediate to advanced level, you are aware that some aspect of a player’s batting average is luck driven along with a function of his skill.  Say we have a player expected to hit .300 that is off to a slow start, perhaps hitting .240.  If asked what the player will hit the rest of the season, there are two usual responses.  The first is he will end the season hitting .300, so he will ht whatever it takes to raise his average to that level, something in the neighborhood of .315-.320.  The second is what’s done is done, he is a .300 hitter, thus he will hit .300 the rest of the way.

Applying my hypothesis, here is how I approach the question.  There is a luck element and a skill element resulting in the .240 average.  The player’s skill is likely lower than normal plus he has incurred some misfortune.  With respect to luck, what is done is done.  I do not expect him to enjoy the necessary level of good luck to compensate the bad luck.  The level of luck going forward should be expected to be neutral.  But, assuming the player’s skills have not unexpectedly declined, I anticipate a stretch where his skills are above normal, so at the end of the day, his skill level is as expected. This does not mean he will end the season at .300, but he will approach it.  Of course, there is a chance his luck does flip and he indeed hits .300, just as there is a chance he remains unlucky, and even though his skills return, his average remains a good bit below .300, yet does increase above .240.

At this point, you may have some questions for me so please allow me to anticipate a couple.  If there is something else you would like me to address, please feel free to comment below of on the message forum and I will do my best to offer a reply.

I bet some of you are wondering what metrics I consider to be skills based, and which oriented in luck.  Long time followers of my work know I am a big fan of looking at strikeouts and walks, for both hitters and pitchers.

For hitters, I rank contact rate on top of the list of skills followed closely by walk rate.  I also consider line drive percent to be a player’s skill which impacts his batting average as well as his HR/FB ratio, which impacts his power.  It should be noted that HR/FB also involves some luck.  I have done some preliminary research that suggests some players may be above league average with respect to base hits on ground balls and others with base hits on fly balls, which raises their BABIP, hence batting average in general and can be considered to be a skill.  That said, until I can dig deeper into this, I will rely on strikeouts, walks, line drives and percentage of fly balls leaving the yard to judge a batter’s skill.  The luck element is primarily his BABIP and as just suggested, some aspect of HR/FB.   If you want to go beyond just homers and batting average, there is luck involved with regards to the timing of the hits, which may result in a high or low number of runs scored and RBI, not to mention stolen base opportunity.

For pitcher’s skills, strikeout rate rules, trailed by walk rate.  Then comes home run rate with the same caveat as above.  Some part of homers allowed is good or bad pitching, some is good or bad luck.  The same can be said for BABIP, but to a lesser degree.  How much a pitcher controls his BABIP is a hot topic, one which I lean to more than most believe, though still not much.   I have long theorized that each pitcher is actually two pitchers – stretch guy and wind-up guy and part of a hurler’s skill set is the level of effectiveness lost when working from the stretch.  My working hypothesis is the better pitchers lose little with runners on base.  I am getting off on a tangent here, but this may help explain why some pitchers always seem to out-perform their peripherals while others under-perform them.  Those pitching better than their peripherals pitch more effectively from the stretch and vice versa.   As you might imagine, the luck elements are just like hitters – BABIP, HR/FB and timing of hits which leads to a high or low LOB%.  LOB% is the generic metric that many of you may know as strand rate.   Strand rate is the specific term coined by Ron Shandler at Baseball HQ and has worked its way into mainstream analysis.  The concept is the percentage of allowed runners crossing the plate for pitchers is a matter of timing and not as much skill.  That is, when hits allowed are random so if a pitcher allows a cluster of hits with runners in scoring position, his LOB% will be low, due to bad luck.  To be fair, pitchers that fan a lot of hitters can have a better than average LOB% as it is really hard to score a run via a strikeout.  Plus, pitchers that limit walks can have a better than normal LOB% as they advance fewer runners to the next base after a walk.

Another question you may have is what makes me think skills will even out over the course of a season?  This is the part I need to research, but my supposition is there are some pitchers, or types of pitchers that give some hitters fits and others that are easier for the batter to pick up.   Say a hitter struggles against lefties and is facing a string of tough southpaws.  Chances are, he will fan a few more times than normal.  But eventually, since the number of lefties faced is pretty consistent from year to year over the course of the season, the batter will see some righties on which he will feast.  Or maybe a batter has difficulty facing soft tosses and in a small sample, the opposing hurlers are all finesse guys.  The results will not be favorable.  Then, the batter faces a string of hard throwers and his skills shine through.  Again, I need to do an analysis on this, but intuitively, my feeling is skills will find their level, with the caveat that a player’s skills may increase or decline.

Finally, you may be wondering is what does this have to do with managing my fantasy baseball team?  As discussed at the opening, with due respect to Pizza Cutter’s study espousing when certain metrics stabilize, we are at the point of the season where we need to make some judgments on some players.  This essay serves as some background to how I will go about evaluating players in future columns.  I will look at the performance and attempt to decipher what is real and what is fluke.  There is nothing I can do about the fluke.  But I can look at the skills portion and make an effort to determine if a player is displaying a new skill level, or if I expect it to return to normal.  Ultimately, this will be the key to the player’s projected performance the rest of the season.

Add comment

Security code

Latest Tweets






Our Authors