Atlanta's Gambling Strategy and Portfolio Risk Minimization
Thinking about the Atlanta draft strategy, I realized that I might have posted the following thoughts on a discussion board but not here, and a search on my blog didn't yield anything interesting other than the search function does not allow the use of the "AND" operand.
Another interesting aspect of the Brave's draft strategy, besides drafting local players, is that they also focus on High School players to a great extent. I haven't read that book on the Braves scouts and their farm system, so perhaps I'm repeating something from there, but I posted (probably on McCovey Chronicles) that successful teams who draft in the later part of the draft (i.e. the non protected draft picks) probably improve their chances at finding a star player vs. just a good or useful player by drafting High School players. I think it works in two ways.
Don't Search for the Quarter Under the Streetlamp
First, there is a self selection/separation that occurs when players go to college. The better ones do well and get drafted by the protected picks (for those who don't recall, I did a study on draft picks that found that after somewhere among the protected picks, the odds of finding even a useful player goes down greatly and good teams always get an unprotected pick) and thus it is much harder for successful teams drafting lower in the draft to find even a useful player by drafting college players, let alone a star player, the odds are greatly against that even for the early unprotected picks.
However, for High School players, while the odds of a big payoff for a team is much lower than a college player who has proven himself, he still has the potential for becoming a star baseball player, presuming the scouts can recognize the skills properly and the team can pay enough/talk them into signing with them instead of going to school.
It is like that old joke about a guy looking for his lost quarter and when the guy helping him ask him where he lost it, he points in the dark and says "over there but the lights better under the lamppost". Searching among the college players is like looking under the lamppost with every other team when you don't have a protected draft pick (and even 11-15 isn't that good either, it is best 1-5 and still better 6-10, but falls off greatly after that, if my memory serves). High School players can still be that diamond in the rough, if your scouts can recognize the raw talent and mental maturity well enough.
Portfolio Diversification of Risk
Unfortunately, I don't recall the correct term for this in investment circles but the second reason it works is because of the properties of how diversification reduces the overall risk for a given investment return, even if you have a portfolio of highly risky investments - that is what made Michael Milken the millionaire (billionaire?) that he is today, applying that principle to junk bonds. Thus, if you draft consistently over time, looking for high potential ballplayers but with the large risk that most will fail, over time you maximize the likelihood that you do find that high potential star ballplayer. I think that is how Atlanta finds their stars in their system.
As people learn in intro business stat classes, any particular investment might have a large risk component, a wide standard deviation that is, but when you diversify your investment portfolio with a large number of these type of risky investments, you reduce the overall risk of your portfolio, particularly the more random the risk is. Thus if you invest in one coin toss, you have a 50% chance of a complete loss, but if you invest in two coin tosses, you only have a 25% chance of a complete loss, 50% of breaking even, 25% chance of two wins, so your risk of losing everything goes down a lot.
However, in the case of drafting ballplayers, you are talking more like lottery level of odds against you winning than a 50-50 coin toss. I think focusing on college players gets you a distribution where you are maximizing the likelihood of finding a useful and perhaps good player, but you greatly reduce the odds of getting a star player (higher bell curve), whereas focusing on high school players gets you a much wider distribution (i.e. greater risk, flatter bell curve) but still leaves you with the greater possibility of finding a star player than you would just drafting college players. And given Atlanta's success at doing this, perhaps it improves their chances of also finding good and useful players, who are also necessary components for any successful team.
Old Man River
But this requires consistent execution of the strategy as well as proper and accurate scouting. And the Braves appear to be doing both, if not to perfection, to very good effect. They had a big crop of players come through for them last season and just used another in the offseason (Marte) to obtain a needed player. They just keep rolling along.
Another interesting aspect of the Brave's draft strategy, besides drafting local players, is that they also focus on High School players to a great extent. I haven't read that book on the Braves scouts and their farm system, so perhaps I'm repeating something from there, but I posted (probably on McCovey Chronicles) that successful teams who draft in the later part of the draft (i.e. the non protected draft picks) probably improve their chances at finding a star player vs. just a good or useful player by drafting High School players. I think it works in two ways.
Don't Search for the Quarter Under the Streetlamp
First, there is a self selection/separation that occurs when players go to college. The better ones do well and get drafted by the protected picks (for those who don't recall, I did a study on draft picks that found that after somewhere among the protected picks, the odds of finding even a useful player goes down greatly and good teams always get an unprotected pick) and thus it is much harder for successful teams drafting lower in the draft to find even a useful player by drafting college players, let alone a star player, the odds are greatly against that even for the early unprotected picks.
However, for High School players, while the odds of a big payoff for a team is much lower than a college player who has proven himself, he still has the potential for becoming a star baseball player, presuming the scouts can recognize the skills properly and the team can pay enough/talk them into signing with them instead of going to school.
It is like that old joke about a guy looking for his lost quarter and when the guy helping him ask him where he lost it, he points in the dark and says "over there but the lights better under the lamppost". Searching among the college players is like looking under the lamppost with every other team when you don't have a protected draft pick (and even 11-15 isn't that good either, it is best 1-5 and still better 6-10, but falls off greatly after that, if my memory serves). High School players can still be that diamond in the rough, if your scouts can recognize the raw talent and mental maturity well enough.
Portfolio Diversification of Risk
Unfortunately, I don't recall the correct term for this in investment circles but the second reason it works is because of the properties of how diversification reduces the overall risk for a given investment return, even if you have a portfolio of highly risky investments - that is what made Michael Milken the millionaire (billionaire?) that he is today, applying that principle to junk bonds. Thus, if you draft consistently over time, looking for high potential ballplayers but with the large risk that most will fail, over time you maximize the likelihood that you do find that high potential star ballplayer. I think that is how Atlanta finds their stars in their system.
As people learn in intro business stat classes, any particular investment might have a large risk component, a wide standard deviation that is, but when you diversify your investment portfolio with a large number of these type of risky investments, you reduce the overall risk of your portfolio, particularly the more random the risk is. Thus if you invest in one coin toss, you have a 50% chance of a complete loss, but if you invest in two coin tosses, you only have a 25% chance of a complete loss, 50% of breaking even, 25% chance of two wins, so your risk of losing everything goes down a lot.
However, in the case of drafting ballplayers, you are talking more like lottery level of odds against you winning than a 50-50 coin toss. I think focusing on college players gets you a distribution where you are maximizing the likelihood of finding a useful and perhaps good player, but you greatly reduce the odds of getting a star player (higher bell curve), whereas focusing on high school players gets you a much wider distribution (i.e. greater risk, flatter bell curve) but still leaves you with the greater possibility of finding a star player than you would just drafting college players. And given Atlanta's success at doing this, perhaps it improves their chances of also finding good and useful players, who are also necessary components for any successful team.
Old Man River
But this requires consistent execution of the strategy as well as proper and accurate scouting. And the Braves appear to be doing both, if not to perfection, to very good effect. They had a big crop of players come through for them last season and just used another in the offseason (Marte) to obtain a needed player. They just keep rolling along.
posted by obsessivegiantscompulsive at 2/10/2006 11:21:00 AM
2 Comments:
I think you're exactly right about the Braves, including the "hometown" theory in your previous post. And the lamppost analogy is fitting - if you're drafting late in round one, why not take a shot on a high-risk/high-reward high school player (assuming such a player is available at your turn)? Or, as the Giants did with Matt Cain, even in the second round. I could support a policy of sure-thing/moderate-ceiling college guy in the first round, followed by a Matt Cain-type of choice in the second round. What do you think?
Well, that sounds good but Matt Cain was a first round pick. There's basically no sure thing for any pick after, say, the 5th pick (my stats showed 43% chance of star or good player in Top 5 picks, 18% chance in picks 6-20; 74% chance of star or good or useful player in top 5; 42% in picks 6-20). So for any even moderately good team (and even the Giants qualified for that in 2005), the pickings are not that great, our 10th pick has 18% chance to be a good or star player, 24% chance to be useful, and 58% chance to either get a cup of coffee in the majors or do nothing at all.
And that's a good point I probably should emphasize more, that you don't pick a High School player just to pick one, but you pick them when you see the high potential there for him.
I'm planning to post again on this but I'll take a shot again: There's nothing wrong with punting a draft pick after the 20th pick. There's been only 2% stars picked in 21-60 and it drops like a stone from there. There's about 9% good players from 21-30, 2% from 31-60. And about 20% useful players from 21-30, about 17% from 31-50, 95 from 51-90 (rough numbers from the splits I've published, I think I'll do a raw dump next time to show how bleak it is).
Throwing away the picks Sabean has done is equivalent to throwing away a Feliz or Tucker, nice players to have but it's not going to kill you because you can get players like them cheap in the free agent market.
Post a Comment
<< Home