Crapshoot Drafting?
by GlassHeart on Jan 7, 2012 8:05 PM PST
Email
PrintA frequent criticism against Billy Beane is that he has drafted poorly, and a common defense is that the draft is pretty much a crapshoot. Here's a look at the numbers, to see if we can shed some light on the argument.
At the suggestion of iglew, I looked at Baseball Reference, which conveniently gave me both the number of players who made it to the majors at each draft position from 1965 to 2011, as well as their total WAR contributions:
In the first graph, we can see a downward trend in the rate at which the players make it to the majors. 87% of players from the first two positions make it to the MLB, and 77% of players from the next two do. The actual percentage would presumably be higher, since these include very recent drafts. It is clearly no crapshoot at these positions, but the trend line crosses 50% at about the end of the first round, so the scouts' performance isn't really something to write home about.
Next, we use WAR to get a sense of what kind of players we're getting. Here, the contributions of the first overall draft pick stands out, at 799.2 WAR to the next best 506.9 WAR of the second overall pick. This would suggest that it's really important to get that first pick, since they average 19.5 WAR to the second pick's 12.4 WAR.
However, note the numerous little peaks above the trend line in the bottom graph. Why do the 10th, 19th, 20th, 22nd, and 30th picks do so much better than expected? Turns out, those positions yielded guys like Mark McGwire (63.1 WAR), Roger Clemens (128.8 WAR), Mike Mussina (74.6 WAR), Craig Biggio (66.2 WAR), and Mike Schmidt (108.3 WAR). What happens if we consider these guys to be outliers, and remove the contributions of the best player from each position?
We can immediately see from the top graph that that a lot of the unusual peaks have disappeared (into the green bars), which means that they represented drafting truly remarkable players who were not recognized early. The second graph, which is the average WAR after discounting the best player of each position, is also telling: beyond the 6th overall pick, the average WAR contributions rarely exceed 5.
The peak representing the first overall pick highlights how important it is to get this slot. They are dramatically better players: 25 of the 41 first overall picks who made it to the majors exceeded 10 WAR, and 16 of them exceeded 20 WAR. Only 18 of 41 second overall picks exceeded 10 WAR, and only 9 exceeded 20 WAR.
Now, I think we can all agree that selecting a Clemens as the 19th pick is pure luck, so it seems fair to exclude these players from consideration, and the conclusion that I'm coming to is that beyond the first five overall picks, there doesn't seem to be an obvious, high-probability way to find star players.
As for the A's, their best picks seem to have been Reggie Jackson (2nd overall, 74.6 WAR) and McGwire (10th overall, 63.1 WAR). Their best 1st overall pick was Rick Monday in 1965, yielding 32.7 WAR, but didn't do so well in the 2nd overall picks (Mark Mulder, Ben Grieve, Pete Broberg). Since 1965, the A's only had ten chances at top 5 overall picks, and these are chances that you weren't supposed to miss, so this had to really hurt.
Thus, I believe today's problem lies with not getting many shots at all at the "can't-miss" slots, and not having gotten lucky with the other slots. I'm inclined to excuse Beane or any other manager for this. I think perhaps people grossly overestimate the advantages of Moneyball, particularly when it comes to drafting young players.
67 comments
|
9 recs |
Do you like this story?
Comments
try hard to get that top 5 pick
like trading away 40% of your rotation and your closer? that’s a good try.
by rollierollieOxenfree on Jan 7, 2012 9:42 PM PST reply actions
But that’s the dilemma, isn’t it? Trade away your good players so that you perform poorly enough to get high draft picks, and fans get mad at you for being a loser. Hang on to them, and you get criticized for drafting poorly. The data I presented above suggests that beyond the first few picks, finding a star is a crapshoot, so if you believe this interpretation, you’re gonna have to allow the GM to do one or the other.
I'd say the Mark Mulder pick turned out pretty well.
Maybe not based on WAR, but he was part of the big 3 and yielded Dan Haren. Not bad at that pick.
Good stuff by the way.
Interesting study.
Yeah, it does sound like I was dissing Mulder...
…which I didn’t actually mean to. I just meant to list the picks, and say that on the average it didn’t turn out well.
But I also need to take that statement back entirely. Out of the second-overall picks, the A’s got a superstar (Reggie Jackson, 74.6 WAR) and Mulder (16.3 WAR). Grieve and Broberg didn’t turn out so well, but since less than half of second-overall picks yield over 10 WAR, the A’s actually did okay.
Greive and Broberg
Great stuff, thanks for the read,
To be fair Broberg did not sign and went back into the draft three years later. But this leads to another question. Wahington, who signed Broberg out of college in 1971 and then I believe never had him pitch in the minors. He made 18 starts in 1971 after being drafted in June. So the question is, was he a bad pick, or was he a good pick the Washington destroyed by not having him get ready in the minors for a year or two?
Which leads to another question. If the 1 pick is injured (Taylor by the Yanks) or drafted because of lower cost (the SD kid that scips my mind) are those picks different? I understand that Taylor was the right guy to go number 1, and then got hurt but the Padres picked Matt Bush (ok, I looked it up) he would sign for less. Pick two was Verlander who went for the same signing bonus as Bush and at pick 12 and pick 15 the two highest signings were Jerad Weaver and Stephen Drew. I would think one of those three were the who the Padres would have taken, if money was not a factor.
Based on the new cap system of the draft, the top 5 picks are going to be even more likely to work out.
As for Grieve, does he turn into a great pick if you add in the WAR of Mark Ellis who he was traded for? The idea being, if the A’s draft a kid at pick one and he flames out, but before he does the A’s trade him for a Hall of Fame player, is it still a bad pick? If you add Grieve’s days with Oakland to Ellis and his time in Oakland you get a 27 WAR.
Anyway fun read, always interesting to look at draft history
Good questions.
Whether or not a prospect was just overrated or destroyed by the drafting team is an excellent question, but ultimately I think something that cannot be answered. Hidden in my graphs is the fact that a star player’s best years may not have been enjoyed by the drafting team at all, which means there’s another skill to either knowing to be patient with the prospect or actually developing him.
I don’t think being traded for Ellis should be a credit to Grieve. The A’s did okay in the end, but the team that got Grieve did not, which means drafting Grieve as second-overall was a bad decision (even for the A’s), unless you believe that you can consistently offload Grieves for Ellises. I think you’ll have to call that a bad pick.
My question is what a "crapshoot" is
I don’t want just a player to make it to the majors. I would expect all the 1st rounds to make it. I think the crapshoot comes in those that become stars. Your Total WAR by Draft Pick graph( with its outlier peaks), tells me that it really is a crapshoot to draft a core player, that we can use to build a team around.
"Trying not to rec a "F**k the Giants" post is like trying not to look at boobs."-anonymous
"i guess i just like beer"-stm
First of all, by the end of the first round, even the low bar of “makes it to MLB” is only reached by 50% of the picks. This was rather shocking to me, as I’d expect as you do that first rounders all (except for freak injuries and such) make it.
I don’t think it’s necessarily a crapshoot when it comes to core players, though. While there are stars who were grossly underrated, they were clearly underrated by the entire league (or they’d have been picked earlier), and so those guys are clearly “crapshoots”. However, 16 of 41 first overall picks turned in at least 20 WARs, so they’re clearly not just randomly selected. I think the answer is that scouts are able to pick out certain stars with great accuracy, but not nearly all of them, and the error margin increases very quickly, such that half of the thirtieth picks don’t even make it to the majors.
So, unless a team somehow finds a way to spot the underrated stars, the first few picks seem still very much worth having.
Cool
I hadn’t noticed this pick when I posted my piece a day later on the front page. I think that there is a definite trend line, obviously the data shows it. But all those #1’s have Bullingtons, Taylors etc mixed in. With Piazzas being in the 60+ rounds too as you write. Great stuff!
Visit my blog the Todd Van Poppel Rookie Card Retirement Plan!!
This is nice, but I have one quibble:
You can’t just throw out the positive outliers in search for a better fit to the trendline. Those outliers were actual picks that produced in the draft slot, which is exactly what you’re looking for here.
Not quite.
What I was looking for is what sort of performance to expect of players from each draft position. Put another way, I’m trying to answer whether Beane should be faulted for not finding stars at these later positions, by first answering what sort of players are typically found at each position.
An extreme example is at position 31, with Greg Maddux contributing 96.9 of a total 122.3 WAR. With him, the average WAR of the position is 8.2. Without him, it’s 1.81. If we further remove Jarrod Washburn’s 26.1 WAR, the average WAR actually goes negative. I believe the right way to interpret this result is to expect replacement-level players at this draft position nearly every time, not 8-WAR types.
Where the contribution of the most positive outlier is more in line with the other players at the same position, they certainly should not be discounted.
Thus, I don’t think that Beane should be faulted for not having picked a 97-WAR player at position 31 in any given year, nor should he be faulted for not having picked an 8-WAR player at position 31, because there actually weren’t any 8-WAR players at that position. The next thing to analyze to answer the question is how Beane performs when he does have one of the first few picks, as well as whether he seems excessively unlucky in later rounds. The problem is that the sample size is rather small.
TL;DR: The taller the green bar is relative to the blue bar, the more the blue bar represents what you should expect to get at that position.
The small sample size does cause problems
However, regardless of your intentions, throwing away outliers is going to skew your results. You come to the conclusion that Beane’s results aren’t much different from anyone else’s, but to get there you’ve thrown out the biggest successes.
One way to help with the sample size problem might be to group draft slots, 1-5, 5-10, 10-15, etc.
But then your clumping would be arbitrary.
You could accomplish a similar thing with a moving average. Of course that won’t work at the first few picks, but it would give a smoother result for the rest. But then you have the question of whether to weight it and how much.
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
Using bWAR...
Mulder produced 17.5 Wins in 5 seasons (including 2 AS caliber seasons per BR’s grading) of pitching for Oakland before being traded.
Grieve produced 5.3 Wins in a little over 3 years for the A’s. And the real problem with Ben was his defense; the A’s almost would have been better sticking him at DH. Of course, rather than sticking him at DH Beane dealt him to Tampa for Cory Lidle (6.6 Wins in 2 years with Oakland) plus Mark Ellis (21.6 Wins in ~8 years) and Johhny Damon (2.7 Wins + the #1 pick that became Nick Swisher) from KC (Angel Berroa and AJ Hinch went to KC as well as part of the 3-team swap).
I’m not trying to shatter your argument GlassHeart (cheezy pun, oh well) just pointing out the need for a bigger picture view at the “success” of a team’s draft strategy. So while it is easy to call Ben Grieve a flop as a #2 overall pick the overall value the A’s got out of the #2 pick in the 1994 draft was immense.
The monster at the end of this blog.
I consciously did not include subsequent trades because, frankly, it’d be too difficult to analyze, but also because I wanted to concentrate on the “eye” aspect. Can teams truly recognize talent in drafts?
If so, I’d expect the WAR plots to be smooth, gradually lowering. I’d expect the rate at which first round picks enter the majors to be nearly 100%. The graph turned out to be more chaotic.
For guys like Grieve (poor poster boy!), the key to me is that the A’s thought that he was the second best player in the draft that year, and he didn’t turn out to meet those high expectations. I don’t want the team’s skills at trading to color that. IOW, because I assume that lopsided trades are not generally repeatable, you shouldn’t get a lot of credit for them to make up for the fact that you drafted poorly.
As I replied to asyouwish33 above, I re-examined the data and my statements on the A’s performance at the second overall position is bogus. Getting Reggie Jackson and Mark Mulder out of four chances is actually pretty good.
I missed the recant on the #2 pick
Recognizing talent doesn’t end once you draft a player, though. Beane traded Grieve (1.4 bWAR left in his career at the time) A.J. Hinch (-0.5 bWAR from 2001 until retirement) and Angel Berroa (3.3 career bWAR) for Lidle, Damon and Ellis.
Ben Grieve was largely considered the best hitter available in the 1994 draft. He had “Talent”. What Beane was able to do was recognize that Grieve’s “talent” wasn’t going to translate as well into production as is hoped from the #2 overall draft pick. Remember, WAR is about production, not talent. Put another way, if Washington were to trade Bryce Harper right now to the Dodgers for Matt Kemp and Clayton Kershaw would you come back 5-7 years from now when Harper hangs ‘em up having never played a big league game and say that Washington picked poorly? They’d have picked a “Talent” that enabled them to produce 10-15 Wins a season.
It is that talent that gives a prospect value beyond however he actually performs as a big league player. Talent acquisition and production has to be veiewed as part of a whole; it involves the draft, free agency, IFAs and trades. Beane could draft 5 1st round out-and-out busts in a row but it won’t hurt Oakland if he deals them for other players/prospects that end up producing for the A’s down the road.
The monster at the end of this blog.
For the scope of this study, though, I find it ok that he limited it to just the drafted player.
Because 1) you might then be faced with an issue of conflating drafting skills with trading skills, or at least blurring that line, and 2) it is too much damn work, and 3) when talking about the expected return of each pick, it seems hard to expect a return via trade.
Basically all #2 picks have high value for at least a few years, because you can always trade them (at least until they have disproven their high value, which usually doesn’t happen with players that good until at least AAA). All #2 picks are generally universally regarded as valuable players by every team until proven otherwise (and incidentally Grieve was pretty good in the majors as well). Thus your analysis, if it includes trades, would reward anyone who drafted a #2 pick and then later decided to trade them, not really rewarding people who recognized the talent (because recognizing a #2 pick level of talent really just takes reading baseball america.
by Billy Frijoles on Jan 9, 2012 3:36 PM PST up reply actions
What I'm trying to say is...
WAR is about production, so using it to grade positive/negative results in a draft position does not necessarily say anything about a team’s ability to judge talent. Sickels writes about this often: Talent is the raw ability. Skill is what leads to production. Talent has to be molded into skill.
Draft selection is also frought with inherent contradiction re: talent and cost of acquisition. It is highly doubtful that the Orioles thought Matt Hobgood was really the 5th best amateur talent in the 2009 draft… but he was the “best” talent on their board who’d agree to sign at their terms. At the other extreme, it’s incredibly unlikely that any team ranked Josh Bell as the 61st best amateur in the 2011 draft but that’s where Pittsburgh took a chance on his bonus demands.
Talent alone did not dictate when those two players were drafted. Therefore a team’s eye for “talent” cannot be judged simply by looking at a WAR comparison for what was produced historically at a certain selection. To have to factor in other variables to create an accurate picture.
The monster at the end of this blog.
that is true, but ultimately judging future production is probably an even more useful skill than judging talent
In terms of creating a winning MLB team. Only because most of the time judging top end talent (not production) is fairly easy (given the somewhat general consensus on who should be the top picks every year), and WAR (i.e. MLB production) is not dependent on getting value from trades.
And yes the other aspect that definitely complicates this study is the change in draft position based on signing bonus demands, but that takes a lot of time to evaluate properly.
by Billy Frijoles on Jan 10, 2012 11:19 AM PST up reply actions
But “talent” in this case is merely a proxy for “production”, because by definition these players have no MLB service to gauge. The objective of the draft is still to generate WAR.
What I’m trying to do is to compartmentalize the various GM skills into at least the following buckets:
1. the ability to select a player who later turns in great production
2. the ability to help drafted players turn talent into production
3. the ability to trade away less productive players for more productive players
Your example is complex, and indeed my simplistic model above would conclude that Washington picked poorly but traded well. But as you suggest, intention obviously counts. If Washington intended (at the time of the draft) to keep Harper, then his eventual non-production is evidence of their poor drafting decision. However, if they drafted him thinking that he’s overrated, and planned all along to dump him, that would certainly be a brilliant thing. I’m not sure we’d ever be able to account for these, though.
I get what you're trying to do, but look at what you said...
The objective of the draft is still to generate WAR.
This is true. However, you then limit yourself by decreeing the only WAR value worth having is that which the draftee himself produces in the Show.
You’re right, we have no way of knowing about a team’s “intention” when they draft a player. But that’s not what I’m shooting for; if you’re going to use WAR as the measuring stick then what is necessary is a GM’s ability to recognize an opportunity to acquire production and to give themselves as many opportunities as possible to do so.
The A’s have struggled to produce a fertile farm system over the past few years not (just) because they’ve struggled to identify amateur talent but because they limited their opportunities to do so. For years they stuck to slot signings and virtually abandoned the Latin American markets. They rarely went after draft & follows when that was still allowed. Combine that with picking in the latter half of the First Round demanded that the A’s excell in the draft or watch the farm system decline. Drafting is not all about recognizing talent… signability is key as well. (Plus some other stuff but let’s keep the discussion to the basics.)
From 2002 – 2007 the A’s had 28 draft picks before the 3rd round. They did not go over slot in any of the follwing rounds. (For the sake of discussion I’m putting the mark of an overslot signing at 2nd round or higher $$$.) Those draft picks produced 10 players that earned prominent roles in the big leagues: starting position players, SP and closers.
From 2008 – 2010 the A’s only had 5 picks before the 3rd round. They went over the suggested slot to sign 3 of those 5 picks. They spent 2nd round or higher money on 6 overslot signings during that time frame. (They also went slightly overslot to sign Chad Lewis in the 4th round of 2010.) It is too early to get an accurate read on the productivity of those draft classes but they illustrate how a team can double it’s opportunity to land meaningful talent. To say nothing of the A’s renewed efforts in Latin America and Asia during these years.
Ironically, it is this last group of picks that will go a long way towards answering your question to Beane’s "eye’ for talent. The 9 players the A’s went overslot on represent “Talent” that Beane decided was worth the extra cost to acquire.
The monster at the end of this blog.
Your points are well-taken, and certainly I didn’t expect the rather flimsy data to be proof that Beane did well or did poorly. I was mainly trying to see if GMs in general were able to pick players who were ultimately productive in the correct order. The various exceptions (like money considerations) I was hoping would be drowned out by the sample size, but I really don’t know if that’s true or not.
Some teams follow a strategy
which leans more toward trading top prospects away in order to obtain established players now (eg, the Phillies). Other teams have a tradition of mostly keeping their own farm players, signing free agents but rarely trading (eg, the Angels).
If your analysis only counts WAR produced for the team that drafted a player, teams which tend to hang on to their own players will look better than they are, and teams which trade a lot will look worse than they are.
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
Overslot
The 9 players the A’s went overslot on represent "Talent" that Beane decided was worth the extra cost to acquire.
And one of those nine is Max Stassi….
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
We have to quit going for the "Safe" College guys
and roll the dice more on the high upside guys. If you’re only going to hit on 50% of your picks, make them high return guys instead of low return safe guys.
I was the first to get Blez autograph. Nebraska native, South Dakota living A's fan
You're missing the obvious
It’s the low return “safe” guys that hit more often, bumping the odds to the 50% margin you suggest. Higher ceiling prospects tend to fail more often; in other words, if you only drafted players with high ceilings you’d have much fewer than 50% of your picks make it.
The real question is, what percentage of high-end guys have to make it for it to make sense for a team to abandon it’s pursuit of “safe” guys?
The monster at the end of this blog.
Depends on your definition of hit
If you mean we get 10 Landon Powells with our drafting philosophy, instead of 2-3 Eric Hosmers, i’ll take the Hosmers every time.
We don’t have, nor seem to draft, impact bats. Instead we draft a safe James Simmons for example who can’t even make our top 40 minors.
I was the first to get Blez autograph. Nebraska native, South Dakota living A's fan
Eric Hosmer's don't grow on trees
You need to have multiple top 5 picks to have the option for a Hosmer type. That said, if you’re in the top 5, even the college guys will have major upside.
Well, Simmons did blow out his shoulder
Injury tends to have a negative impact on any prospect’s ranking.
As for the definition of hit… it’s your definition we’re playing with. You said “If you’re only going to hit on 50%” so if by that you meant the Landon Powells of the world then I wish you had been more clear in your original comment.
The monster at the end of this blog.
It's probably the definition in the original post, as he used the 50% number which was the % of players that make it to the majors
Which includes both Landon Powells and Eric Hosmers, so unfortunately not that great of a definition. I guess high-risk/high-reward types would mean less middling players and more complete busts, but maybe one boom.
by Billy Frijoles on Jan 10, 2012 12:11 PM PST up reply actions
This is more what I was talking about
I would rather not land 10 guys who just make it to the majors, like powell, quintanilla, etc, an rather takle 20 shots (to land the 10 above) at guys who could turn out to be special, and only end up with 2 or 3, but high impact guys
IE -
draft 10 Landon Powell types, and hope 5 make the majors as backups
draft 10 high upside high risk guys, with the knowledge that maybe only 2 or 3 make a big impact
I was the first to get Blez autograph. Nebraska native, South Dakota living A's fan
Landon Powell was a 1st round pick
Who before a knee injury and then liver problems was projected to be a starting caliber Catcher in the Show.
The monster at the end of this blog.
So, instead of joining the priesthood, he just destroyed his body?
"Trying not to rec a "F**k the Giants" post is like trying not to look at boobs."-anonymous
"i guess i just like beer"-stm
I think he's secretly a yakuza
"And Julio Franco is batting right-handed!" -- Wayne Hagin, A's radio play-by-play, mid-80s
Not very considerate of him,
when we take into account his high pick and all. :) This really does go to show what a crapshoot it really is, both in the drafting procedure, and the success of the individual player.
"Trying not to rec a "F**k the Giants" post is like trying not to look at boobs."-anonymous
"i guess i just like beer"-stm
Well, the health "skill," as much as there is one, is really poorly understood.
So that does add a huge crapshoot element to the whole thing.
Get out the time-fracture wickets, Hobbes! We're gonna play Calvinball!
I was just musing that Powell and Desme, even though they have differing situations,
have affected the A’s in a very dramatic fashion.
"Trying not to rec a "F**k the Giants" post is like trying not to look at boobs."-anonymous
"i guess i just like beer"-stm
Certainly have.
Desme is the very definition of unanticipatable (it’s a word, swear) bad luck.
Get out the time-fracture wickets, Hobbes! We're gonna play Calvinball!
Great visual look into the first round
Definitely rec’d. Most striking to me was the large drop-off in WAR right after that spike at pick 22.
I’m mildly surprised to see the solid downward trend in appearance percentage. It’s a really weird metric to measure players with. Guys on bad teams may make rosters even when they’ve no business being there. Then there’s September call-ups, and there are all sorts of motivations for giving guys a cup of coffee.
All told it’s a challenging data set and there’s all kinds of complexity involved. But you gotta start somewhere and I applaud your effort.
Ending on a frivolous note, I’m quite excited by that big WAR spike at pick #10. It’s obviously not statistically significant, but even so seeing so many #10s do well gives me more hope for Michael Choice (10th pick in 2010).
Wow, really cool analysis
It makes me wonder if the same trends hold true in more recent drafts. My gut feeling is that teams have gotten more successful at drafting in recent years due to increased money and focus on acquiring amateur talent as well as technological improvements in travel and communication which allows scouts to get more information on potential draftees.
Probably no way to tell if my hunch has any validity because narrowing the data to only look at certain eras would lead to really small sample sizes.
This is really cool. I don't see how it proves anything about a "crapshoot" though.
You haven’t shown that there is no skill involved in drafting. It’s like drilling for oil. Most oil wells are duds, but that doesn’t mean that geological research doesn’t improve your odds, nor that some are better at it than others.
If you really want to show that drafting is all luck, you need to show that there’s no linear correlation in drafting, i.e. that there are no GMs (cough Friedman, Moore cough) that consistently draft better than other GMs, and others who are worse (cough Littlefield cough).
I find it incredibly unlikely that there’s no skill involved in drafting.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 12:32 AM PST reply actions
Was reading
the Boston Globe and they were talking about compensation picks made by the Sox under Theo. Was an interesting list. Even though there were a few duds in there, there were some impressive picks, such as Ellsbury and Papelbon. I’ll see if they posted the column online and pass on the link later today.
I think a GM by GM analysis would be interesting – many of the hits by the Sox were mid to late first round, around where the A’s have been picking. Did they pay over slot? Did they just have a good eye for talent?
The Rays picked Matt Moore in the 8th round and Desmond Jennings in the 10th.
You could argue that those two picks were better than any A’s 1st round pick since Zito, depending on what you think of Swisher and Weeks.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 9:06 AM PST up reply actions
Matt Moore in the 8th may end up being the equivalent
of Rich Harden in the 17th. Might depend on whether Moore’s arm keeps falling off.
"And Julio Franco is batting right-handed!" -- Wayne Hagin, A's radio play-by-play, mid-80s
Could very well be, but he's worth a lot today.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 2:19 PM PST up reply actions
About what Harden was worth before he got hurt, probably
That is, using career WAR isn’t a great way to compare A’s picks from 10 years ago with a guy who just made it to the majors last year.
"And Julio Franco is batting right-handed!" -- Wayne Hagin, A's radio play-by-play, mid-80s
Right, he probably is worth about what Harden was worth in 2003.
But you can still use prospect value thing, no? Of course I’d consider Moore better than a mere “A” prospect.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 3:04 PM PST up reply actions
Well, I don’t think I was trying to show that drafting is all luck. It’s clear that the GMs consistently know who the very best handful of players are, which is why the first few picks are so typically successful. What I’m interested in finding out is whether they are similarly prescient with the rest of the pack (whether the graphs come out smoothly descending), and if so, whether Beane’s picks deviate. The chaos doesn’t say Beane is very good (because I really barely looked at Beane specifically), but neither does it show that he’s unusually bad as some have claimed.
The other thing I wanted to see was if there was a “cliff” after which prospects suddenly become much harder to evaluate accurately.
So you're saying it's inconclusive?
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 2:20 PM PST up reply actions
Absolutely.
This one is merely a cursory search for obvious patterns. I would not expect it to reveal anything deep.
One thing that did stick out, however, is how important the first few picks are, unless your GM really could consistently pick great players from later positions.
You don't have to "consistently" pick great players from later positions. You just
benefit from doing it occasionally or even once as opposed to never. Nobody consistently picks great players at any draft position. There aren’t that many of them.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 3:02 PM PST up reply actions
But that’s exactly what some people seem to be demanding of Beane, which brings us to why I decided to look at the numbers. :)
I'd settle for one great player every 10 years and a good one every 2 years
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 10, 2012 8:35 PM PST up reply actions
Well, depending how strictly you define "good" and "great",
I think the A’s meet that standard.
Since the 1980 draft, here’s what I come up with for “good”:
1981 = Mike Gallego
1981 = Mickey Tettleton
1982 = Jose Canseco
1983 = Terry Steinbach
1984 = Mark McGwire
1985 = Walt Weiss
1986 = Kevin Tapani
1986 = Rod Beck
1987 = Scott Brosius
1992 = Jason Giambi
1993 = Scott Spiezio
1996 = Eric Chavez
1997 = Tim Hudson
1998 = Mark Mulder
1998 = Eric Byrnes
1999 = Barry Zito
1999 = Ryan Ludwick
2000 = Rich Harden
2002 = Nick Swisher
2002 = Joe Blanton
2003 = Andre Ethier
2004 = Huston Street
2004 = Kurt Suzuki
2004 = Dallas Braden
2006 = Trevor Cahill
2006 = Andrew Bailey
2008 = Jemile Weeks?
That’s close to one per year (with a strange dry spell between 1987 and 1992). Even if you’re pickier than me and weed out a bunch of the borderline ones, it still makes two per year easily.
So how about “great” players? Again, it depends how picky you are. I would say Canseco (1982), McGwire (1984), Giambi (1992) and Hudson (1997) for sure, and possibly also Steinbach (1983), Chavez (1996), Zito (1999) and Swisher (2002). That would make us overdue for the current decade’s great player, but averaging well over one per decade before that.
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
by iglew on Jan 10, 2012 10:10 PM PST up reply actions 1 recs
Awesome list.
Just focusing on the Beane years,
I do think Swisher and Andre Ethier would be considered great players(although if he had stayed in the A’s system who knows if he would have achieved the success that he has). Probably Andrew Bailey (although his injuries ding him). I hope Michael Choice can be another one.
Although my definition of great is really just multiple all-star. Some people might have a higher bar, like borderline HoF.
by Billy Frijoles on Jan 11, 2012 11:15 AM PST up reply actions
Very nice! But I was thinking of something a bit different.
The top 5 AL teams in 2011 — Rangers, Yankees, Rays, Tigers, Red Sox — averaged about 32 WAR from their top 6 players, or about 5.5 WAR per player. The range was from 30.4 (Red Sox) to 33.6 (Rangers).
If you want to be a World Series contender in this league in this day and age, you basically need at least 1 great player (7+ WAR), and at least 5 good players (4+ WAR). If you just meet the bare minimum (One 7 WAR guy and five 4 WAR guys), you’re at 27 WAR and your not a playoff contender in the AL without some luck.
If you’re going to do this without big money free agents, you basically have to do it in the draft, IFA or trades. The early 2000s A’s had this in Giambi, Tejada, Chavez, Hudson, Mulder and Zito, and the 2011 Rangers (Kinsler, Beltre, Napoli, Hamilton, Andrus, Wilson) and Rays (Zobrist, Longoria, Upton, Price, Shields, Joyce) are doing it, so it can be done.
Looking at these 3 sextets, you’ve got 11 draftees, 1 IFA and 5 trade acquisitions and one semi-expensive FA.
So if you’re going to use the draft for 4 of your six core players, you’ve got to get one great player (7+ WAR) and 3 good ones (4+ WAR). Since 1998, the A’s have drafted 0 players who have ever managed a 7+ WAR season, and only 3 (Mulder, Zito, Harden) who have had more than one 4+ WAR season.
The list is good for proving that the A’s have drafted a fair number of averagish players, but it also shows that their drafting has been nowhere near good enough to compete in the AL as it currently stands.
The usual argument is that the Rays drafted early a lot and therefore got Upton, Longoria and Price, but they also got Zobrist and Shields very cheaply, and of course of the Rangers’ sextet only Andrus was acquired in a trade for Teixeira (a top 5 pick). Even Andrus was only a B+ prospect at the time of the trade, however and was probably available for a lot less than Teixeira.
tko bira, masturbira -- Croatian proverb quoted by elcroata
by WaddellCanseco on Jan 12, 2012 3:40 PM PST up reply actions
The Padres got Jason Bartlett in the 13th round
Russell Martin was a 17th rounder…
There are so many late round picks that it’s probably lucky that the Rays made two really great picks. If they loved Moore and Jennings so much, why didn’t they draft them in the 4th or 5th rounds, where the marginal signing bonus demands would not have mattered? Why risk another team taking them?
That doesn't necessarily mean it's all luck though
They could have some inherent advantage in judgement that helps them pick guys with a higher chance of succeeding. In other words, every one of their picks had a higher than average chance of being good (due to this hypothetical ability the Rays’ FO has), and these two were the late-round ones that happened to pan out. I’m not saying this is the case, just that it’s plausible that it’s not pure luck.
I’d guess that sample size issues would make any analysis of different teams’ ability to predict late round hits virtually useless. That’s just a guess, though.
But ideally you want the level of success to match the draft position, because in general you don’t know what other GMs think of the guy, and if you like him you don’t want to wait too long to grab him. Late round picks who turn out great should, by default, be attributed to luck until we have some evidence that it’s actually cunning.
Also, if your 13th-round pick grossly outperforms your 3rd-round pick, I would probably also attribute that to luck until disproven.
There's a distinction between blind luck and playing the odds.
It’s a perfectly reasonable strategy to pick safer players ahead of super risky players with high-ish ceilings. If one of those super risky players does click and outperforms the earlier drafted safer player, it doesn’t mean you fell into that player by chance. It just means that you weighted him where your risk-reward analysis suggested and the gamble paid off. Just because a player works out down the line doesn’t mean it would have been a smart move at draft time to take him in an earlier round.
Get out the time-fracture wickets, Hobbes! We're gonna play Calvinball!
Remember, laserbeams is talking about an “inherent advantage.”
I’m saying if a GM did have an inherent advantage, he’d want to draft the risky-to-everybody-else players earlier than you might expect, to avoid losing him. The better this “inherent advantage” at foretelling production, the more you’ll want to use an early pick on him…
…unless you also have very good knowledge of how other GMs rate him, in which case you can just wait until the round before the next guy would’ve picked him. The better you can gauge this, the later you can wait to draft him.
IOW, the first half is as if I gave you a lifetime WAR listing of historical baseball players from the year 2050. This list should make you want to draft Albert Pujols in the first round. The second half is as if I also gave you the actual draft order, so you can see that you could land Pujols anytime before round 13 to beat the Cards to the punch. Thus, based on the fact that Pujols was signed in the 13th round, the GM either had a very good idea how he’d turn out and a very good idea that he’s grossly underrated by other GMs, or it’s more luck than skill.
The former possibility is, of course, extraordinary, and would require extraordinary proof.
…
By the way, the other obvious possibility is that a team has better teachers and trainers, so their prospects achieve their potential more often and get injured less. The GM may not be any better at picking them, but can simply be better at growing the ones he does pick.
GlassHeart:
The convention on AN is to use subject lines for comments. This facilitates various database-related access to your comments, such as searching. It also gives more viewing options for those who are accessing AN on something other than a standard browser.
Unless you have some personal or stylistic reason for wanting to avoid them, I would recommend you get in the habit of using subject lines for your comments.
Thanks.
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
I'll try harder.
Usually I omit the title if I can’t quite summarize the rest of the post with it. :)
Just stick the first sentence in there.
That’s what everyone else does.
Baseball is a stupid-making enterprise in that nobody wants to be singled out or say something dumb. —Michael Lewis
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Athletics Nation
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
Baseball Nation Recent Stories
In Association With
Big League Stew - MLB - Yahoo! Sports
- Barry Bonds thinks it will be ‘very sad’ if he doesn’t get into the Hall on his first try
- Chris Johnson’s glove webbing had a Spidey sense at Coors Field — for one play, at least
- Jonathan Lucroy breaks hand after suitcase falls during search for sock
- The Juice: Long Adam Dunn homer punctuates Chris Sale’s 15 Ks, Mark Trumbo goes deep after Jered Weaver goes down for Angels
- Leyland begs media to make umpires ‘accountable’ for bad calls (VIDEO)
