"Those who cannot remember the past are condemned to repeat it."
--George Santayana, "Life of Reason I"
Perhaps the biggest topic of discussion among Astros blogdom is the upcoming Rule 4 draft in which MLB clubs select players from High School or College to fill the ranks of their minor league systems. The Astros wield the dubious honor of selecting first – a backhanded award for finishing dead last in 2012.
There has been much behind-the-scenes discussion among TCB writers over who the Astros should select, as there presently is no player who is obviously the best candidate. Compounding the debate is the fact that the Astros’ brain trust is as out-of-the-box as one can find among major league front offices (offici?) and we learned last season that expecting them to do the expected is futile.
I endeavored to quantify some of the questions commonly asked about the draft by diving into historical data and swinging my Magic Sword (MS) Excelibur.
Chances of making the majors:
The first and perhaps easiest question to answer is – what is the likelihood of a certain type of drafted player even sniffing the major leagues? The four major categories I investigated were:
- HS P – Pitchers drafted out of High School
- COL P – Pitchers drafted out of College
- HS H – Hitters/Position Players drafted out of High School
- COL H – Hitters/Position Players drafted out of College
At this point, I expect somebody to suggest, "Wouldn’t this be better if you broke it down by position instead of just by hitter or pitcher?" The answer is yes, it would, if I had enough data. But alas I fear that if I break it down any further than I have, the sample sizes of each type would be so small as to be completely meaningless instead of just possibly meaningless. For example, how many catchers were drafted in the Top 10 over those fifteen seasons? What about second basemen? And how do I account for position changes? Better left alone, trust me.
Following in the footsteps of my forebears who examined the draft, I chose a sample of fifteen years’ worth of data – specifically, I looked separately at the Top 50 picks and the Top 10 picks for all seasons between 1993 and 2007. I judged 2007 to be the most recent year where every draft pick has been given a reasonable chance to reach The Show. Heck, it’s been six years, that’s plenty of time! Here are the totals that fit that criterion:
Finding out who did not make the majors was easy, with the help of Fangraphs. I exported a list of all players that played from 1993 forward and who recorded a WAR (Wins Above Replacement – the value of a player compared to an average replacement player, quantified by number of Wins added or subtracted to his team). If a player didn’t show up on that list, he never reached the majors. With a few swings of Excelibur, I carved the data into the following chart:
Given a more positive spin, here was the rate at which the 616 draft picks in my sample reached the majors at all:
WAR per Season by Pick Type, per Pick.
Another forbear looked at value of players picked at certain draft positions. I decided to duplicate his research with some tweaks because I needed the raw data for my next tricks (sorry, you’ll have to wait for my next post to see them), and because I wanted to break it down by player type.
Because players have varying levels of longevity, I opted not to examine career total WAR. Instead, I looked at each player group's average Career WAR per Season. The samples start to get questionably small, but it is still worth examining to see what occurred after past drafts. Players not reaching the majors were excluded from this exercise, as there’s no way of knowing what their WAR would have been had they succeeded.
Graphing WAR/Season by pick number and type gave the following results:
Fangraphs also has a nifty stat called "Dollars", explained here. Dollars converts the wins into cold, hard cash, to give an approximation of the dollar value of a player based on his performance. As with WAR, I modified this to Dollars/Season and graphed it by pick number and type.
Obviously, because Dollars are based on Wins, the two graphs telegraph each other, with only slight variation.
I will be using the data presented above in a later post to calculate even more odds of a player being successful based on draft type. In the meantime, here are some observations about just what has been presented so far.
- Although High School pitchers were the least likely to even reach the majors in the first place, those who did reach the majors were more likely than any other prospect type to perform at a high level. It's interesting to point out that during the time period sampled, whether a Top-10 drafted High School pitcher made the majors or not was basically a coin-flip.
- Note the amazing rate of college hitters reaching the majors. And they performed well also; better than all players except High School pitchers, even with a sample size of over 100 players. It’s safe to say that drafting College Hitters carries the least risk, but sacrifice in ceiling is not as extreme as some people claim, if it even exists.
- College pitchers were very likely to reach the majors, particularly if drafted among the Top 10. However, their performance was among the lowest of the four categories, and was virtually tied with the productivity of High School Hitters. On average, college pitchers in this sample had a much lower ceiling than college hitters or high school pitchers.
- The only college hitter drafted in the Top 10 to not make the majors during this fifteen-season sample was Chad Green, drafted 8th overall by the Brewers in 1996. Offensively, he was a disaster in the high minors and never sniffed the bigs. He retired in 2005 at the age of 30.
- Despite a penchant for many clubs to draft "toolsy" High School position players near the top of the draft, one in four high school hitters never made the majors. Among those who did, High School hitters posted the lowest average WAR and Dollar value among all of the four player types. This data included Alex Rodriguez, who with his 5.85 WAR/Season was bound to skew the trendline's slope by a decent amount all by himself.
Several things are worth noting about my analysis, chiefly that I am no student of probability theory.
- Standard sample size cautions apply about making projections based on the data shown. It’s merely illustrative of what occurred in the past, and might be predictive of future drafts.
- I would have preferred to look at WAR for the just the players’ first three MLB years (when they are under team control and pre-arbitration). This is what clubs are really drafting for. Once a player reaches arbitration, his value changes from "top draft pick" to "guy who makes some serious bank". Risk/Reward analyses regarding the draft don’t apply anymore. However, I didn't have the resources to manually pull seasonal WAR data for over 600 dudes.
- I am making no claims as to WHY players fail to reach majors. Injury, lack of development, lack of maturity, or whatever could all be possible causes. I did some work looking at risk of landing on 60-day DL or its minor league equivalent, but the available data is sketchy, and frankly it's irrelevant to the grander scheme of whether or not a player of a certain type makes the majors or succeeds, other than to explain why.
My next post will expand the discussion to other probabilities based on the same fifteen-year sample, including the odds of a draft pick being a Top 10% performer by WAR, the odds of reaching a threshold of 1,000 plate appearances or 300 innings pitched, and the odds of a player recouping the cost of his salary and bonus over the first few years of his rookie contract.
I will follow that by applying all of these odds to a list of the Top 20 draft prospects for 2013, assembled by the prospect gurus here at TCB by my request with weighted scores based on typical scouting categories. Whew!
At the very least, all this should prompt discussion, and as always I’m open to questions, comments, suggestions, or even snide remarks.