2012 MLB Farm System Ranking Based on Sickels' Grading System
by Frag on Jan 1, 2012 5:22 PM EST
Email
Print
Hello all!
While looking through John Sickels' top 20 prospect rankings, I was interested in coming up with a simple way to rank farm systems based on his prospect grades. Plus, I just wanted to do something while passing the time. So I thought, "Why not try something similar to the GPA system?" More after the jump.
I devised a grading system for prospects similar to the Canadian GPA system that I'm used to, though you can try other grading methods as well, which goes as follows:
A+: 6.00*
A: 5.00
A-: 4.17
B+: 3.50
B: 3.00
B-: 2.67
C+: 2.33
C: 2.00
So on and so forth.
*John rarely (most likely never) gives out the A+ grade
Edit: One commenter suggested a simple rate down from A+ to B, so I did just that. Between A+ and B, the difference between each grade decreases at a rate of ~0.17 per grade. It is arbitrary, I know, but I think this may address the issue regarding the undervaluing of higher grades in a simple manner. It does not, however, address the differences in value between pitchers and hitters.
From there, I simply gave each prospect a specific GPA. For instance, Jesus Montero was a Grade A prospect according to John, so he gets a 5.00. Borderline prospects (Between grades) were given an average (ie. a B+, borderline A- prospect was given a 3.84). Once all of the prospects within a team's top 20 list is graded, an average of the grades for each team was done. The teams were then ranked based on these averages. These were the results I got so far (this will be updated):
1. Toronto Blue Jays - 2.9850
2. San Diego Padres - 2.9505
3. St. Louis Cardinals - 2.9325
4. Atlanta Braves - 2.8510
5. Texas Rangers - 2.8415
6. Seattle Mariners - 2.8380
7. Tampa Bay Rays - 2.8130
8. Arizona Diamondbacks - 2.8085
9. Kansas City Royals - 2.7500
10. Pittsburgh Pirates - 2.7365
11. Boston Red Sox - 2.7340
12. Colorado Rockies - 2.6835
13. Oakland Athletics - 2.6820
14. Minnesota Twins - 2.6545
15. Baltimore Orioles - 2.6365
16. Los Angeles Dodgers - 2.6205
17. Los Angeles Angels - 2.6000
18. Washington Nationals - 2.5965
19. New York Yankees - 2.5905
20. New York Mets - 2.5825
21. Chicago Cubs - 2.5655
22. Detroit Tigers - 2.5485
23. Cincinnati Reds - 2.4905
24. Philadelphia Phillies - 2.4830
25. Miami Marlins - 2.4830
26. Cleveland Indians - 2.4805
27. San Francisco Giants - 2.4665
28. Milwaukee Brewers - 2.4245
29. Houston Astros - 2.3990
30. Chicago White Sox - 2.2985
One flaw of this grading system includes the way I graded borderline prospects (eg. I graded B+/borderline A- prospects and A-/borderline B+ prospects the same at 3.84). As well, I may have missed some trades that involved the exchange of top 20 prospects. Finally, ties were graded equally, even though John likely have had good reasons to rank one B+ prospect higher than another B+ prospect.
Critiques would be greatly appreciated.
125 comments
|
2 recs |
Do you like this story?
Comments
Has John said Grandal/Alonso were B+/borderline A-?
We likely won’t know until he goes back and adds new guys to systems after he’s done with every team, but I feel like they’d both be flat B+’s.
He wrote them as B+/A- prospects
I will update if he changes them, though.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
At least
In the comment section of the trade thread.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
yeah, you're right
I went back and checked the “Prospects in the Mat Latos trade” thread
by walnut falcons on Jan 1, 2012 5:42 PM EST up reply actions
Agreed
I’ll update this once things are set.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Biggest Problem I see
Is that it’s assuming linear differences in grade levels, which is not reflective of the actual situation at all.
In reality, the difference between a grade A prospect and a grade B prospects is bigger than a grade B and grade C, yet it is equal in the these rankings. So basically, it’s going to drastically undervalue systems with one stud player (Washington, Anaheim, as examples), and overvalue systems with a lot of depth, but lacking the truly scarce studs (Toronto, San Diego)
I had been thinking trying something like this, but the most important thing, for me at least, is to get the weights right, or the aggregate ranking don’t mean much. I recall seeing last offseason average $ values associated with each prospect grades, this would probably be best weighting to use.
I recall seeing last offseason average $ values associated with each prospect grades, this would probably be best weighting to use.
I didn’t know about this. Can you provide a link?
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Wish I could
I’ve tried searching for it, and can’t remember enough particulars…it may not have been here, but on a related SB Nation site.
Hopefully someone will remember it, and can provide the link
And
I feel like I no longer need to do this now. Haha
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
I wouldn't put too much emphasis on that
The data for pitchers is incredibly inaccurate as its over a decade old and was calculated on an especially bad time period for pitchers in general (other than Mussina and Pedro, there weren’t many good pitchers debuting in the 90s). Also, something like 1/3 of the data points used (can’t remember the exact number off the top of my head) are 4 pitchers that got ranked a total of 11 times and ended up all busting, so that kind of skews things.
I have used
This system in the past two seasons (and will again do it this year).
I posted it here last season, but it seems to be gone from the fanposts now.
MjwW
Provided a link to it above.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Can we please adjust the pitching prospect values?
Like I mentioned above, the data that’s based on has little relevance to the guys who are in the minor leagues right now. It completely overvalues hitters and undervalues pitchers. If those numbers were right, don’t you think smart teams would always take hitters over pitchers in trades? You don’t think guys like Billy Beane have seen those numbers? If they were accurate, he wouldn’t be targeting mostly pitchers in the deals he’s made this offseason.
Scott McKinney did some related research
And his data is available here (I’m guessing this has been discussed here, but before I joined – sorry if this is rehashing old ground). Would using this data set help to establish better values (could even look at adding 2004 probably since another year has passed)? Would it be better to use say 1994-2003? Or it a problem with Baseball America’s rankings (I believe I’ve seen this suggested)?
I guess what I’m getting at is, what would be your suggestion for the right way to go about getting “better” values?
It might be better, but it would still be subject to quite a few flaws
The data your looking at is always going to be dated because of the lag to see how guys perform in the bigs.
I'm a little confused
Your initial criticism was the the data was flawed because it was centred on a time period when pitching prospects were not developed. I understand this, and the solution seems to me to be incorporate more data, perhaps on either end or both (did BA have top-100 lists prior to 1990?).
Another criticism I have heard is that BA didn’t do a very good evaluating pitching prospects, and so this creates a bias. The only solution here would be a different source, and there’s probably not another public source that could be used (that I’m aware of).
The data your looking at is always going to be dated because of the lag to see how guys perform in the bigs.This, however, seems to be a new criticism of the methodology used to determine the value. Now, I would prefer to use discounted WAR to measure value, rather than $, since $ inflates. Certainly the way you discount matters, but I thought the way it was handled by Wang and subsequent studies was a solid approach. So I’d be curious as to your criticism in this regard (and I’m not trying to be argumentative, I’m interested in conceptualizing improvements to get better info, and then possibily actually gathering data and crunching numbers)
Just look at it this way
Even if you to extend the research to 2004 or 2005, you’re still looking at data that is over 5 years old. In the intervening period, BA has gotten better at evaluating guys and teams seem to be getting better at developing young pitchers (or there is just that much more pitching talent in general). I mean Justin Verlander wouldn’t even be included in this sort of study because he’s come up too recently.
Alright
So this would lead me to a couple observations
1) Even if the data is 5 years old, unless the rate of improvement by evaluators was very quick, then the conclusions should still hold some validity. Also, if over time the rankings are becoming better, we should be able to split up the data and test for this, and then possibly control for it. The other thing is if post-2005, there had been fundamental changes to the way prospects are evaluated such that pre-2004 evaluations were so different as to be useless. Now, there’s more info out there, stats, etc, but I don’t think that’s the case.
2) Even if there is more pitching talent, WAR is basically a zero sum game since you cannot create more wins. More talent raises the replacement level pitcher. The more important question is whether it changes the distribution such that the results of 1990-2004 date or so wouldn’t matter. There’s compelling evidence to suggest that the variance of talent in MLB players has decreased over the last century, meaning that truly exceptional players (relative to their peers) are more valuable. This could suggest old data wouldn’t apply, but the trend is naturally deaccelerating, and we’re still talking less than 2o years total
WAR as a whole is a zero sum game
But its not when you’re only looking at a portion of all MLB players. Look no further than the 90s with regards to pitching. A huge amount of pitcher WAR was eaten up by some talented guys who came up in the 80s.
As far as the data being 5 years old, while it might not be that inaccurate, its the most recent data your looking at. You’re also including quite a bit of data that is over a decade old, and prospecting has come a long way in the past decade.
Fair enough
I personally feel though that it can potentially have more utility than you’re suggesting, at least in informing how we think about trades.
Teams clearly ignore it when it comes to trades
I think that should tell you all you need to know about how well it does at informing us of how we think about trades. I mean it does seem like a solid estimate for hitters, but I think we all realize how flawed it is when it comes to pitchers…I mean unless people would rather trade for a 26-50 pitching prospect than a top 10 one.
Right
No, I meant the framework generally. I acknowledge the pitcher value is off for the perfectly valid reasons you’ve suggested, but if you add some data to smooth it, then I don’t well why it can’t work better, as it does for hitter.
(In other words, I see it as a sample size problem, but a problem where the framework is busted. My understanding was, even beyond the pitcher values that resulted, you didn’t see the framework as very useful)
I think it could have its uses
But even at its best I would see it only as the roughest of estimates. Maybe if we had decades worth of data with multiple ranking systems I could see myself becoming pretty comfortable with the numbers the framework would spit out.
Framework
The framework is great. Even the sample size isn’t much of an issue outside the top 10.
Hitters 11-50=$21.5
Pitchers 11-50=$14.2
Ten years worth of prospects is going to be close to 200 of each. Those numbers aren’t exact, but that’s a big enough sample to feel highly confident.
For the next group it’s going to be 250 of each, then even more for the B group.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
"teams clearly ignore it when it comes to trades"
you do realize, that trades are basically auctions? and that the tables provide the expected value? and that at auctions, the highest bidder pretty much universally overpays at above the expected value?
You realize you just helped prove my point, right?
If teams were using the chart and taking the highest bid, then how come the highest bid rarely includes the significantly more valuable hitting prospects? And why did one of the GMs most likely to be aware of this research (Beane) always end up accepting packages built around pitching instead of hitting?
because
he trusts his own system more than he trusts BA’s system at ranking pitchers?
obviously there are limitations. you’re paying $30 a month for a data set and opinions, whereas actual GM’s are paying thousands and tens of thousands of dollars a month for to try to calculate the same goal we are.
for a stand-in though, this is sufficient. your criticism, as others have mentioned, is simply a matter of degree. it doesn’t make the data/valuation worthless. just take the data for what you think it’s worth, and make your own mental adjustments as necessary.
i personally think BA is terrible at judging pitching talent, and constantly overrates pitchers with “good stuff” and no strikeouts to back it up, which leads to a lower expected value, than if those pitchers had been dropped below the Top 50, or rearranged in the Top 11-50 range.
use the chart as a baseline, and let everyone else make individualized adjustments in their mind as necessary. i don’t see why there’s anything wrong with that.
The system has little to do with it
The results of Wang’s study shows that top 100 hitters are way more valuable as a group than top 100 pitchers. Unless you think Beane and BA basically don’t share much of any of the same pitchers in their top 100, then it should apply to either. If hitters are more valuable, he should be targeting them. Instead we’ve seem him again and again target pitchers.
there's also the factor...
that its probably easier to identify hitters so you wont need as many to get a couple of good ones. whereas you might have a group of 10 good pitching prospects and only get a couple of decent ones out of the bunch.
you can get decent hitters without giving up a ton…beane did it with ddj and willingham last offseason. how much would a carlos quentin level pitcher cost on the trade market? much more than the actual carlos quentin did.
Fire Everyone
by billybeingbilly on Jan 2, 2012 2:25 AM EST up reply actions
Economics
You produce what it is easiest to produce and buy the stuff that is harder to produce. If its tougher to identify good pitchers you don’t trade for pitching prospects. You trade for the hitting prospects you can evaluate correctly and once its value goes up you can flip it for proven pitching.
Don't think it's that simple in practice
The reasons there’s generally a shortage of good pitching is that pitchers are much more likely to suffer catastrophic injuries. Which contributes to the relative supply/demand balance. Hitters are much less likely to have this, and are more consistent contributers. I would think this is part of the reason pitcher value is so much less as well
How does this disagree with what I said?
Why are you trading for the riskiest assets possible? Why not trade for the far safer hitters, let them build value, and then deal them for proven pitching? Why trade proven pitching in the first place, just to get risky assets in pitching prospects in return?
b/c proven pitchers cost a ton of money....
even if it’s in arbitration
Fire Everyone
by billybeingbilly on Jan 2, 2012 2:47 AM EST up reply actions
Because the market is dynamic. If, for example, you are better at developing pitchers can average (like the Rays seem to be for instance), then you trade for pitching, and create more value by developing it, value which you can then trade. This is essentially a theory of comparative advantage.
Or maybe you have a ballpark that makes your pitching look better than it is, and some teams haven’t caught on, so you can “create” value (or at least gain it) this way.
I'm done for the night
But I simply don’t think those things would come close to making up for the relatively large difference in terms of Wang’s values. If you disagree that’s fine. It just feels like everyone keeps trying to defend the numbers based largely on the fact that it is a cool concept. I just don’t see the case for pitchers being so much less valuable based on what actually happens in reality. If the difference was as stark as the picture painted by Wang’s numbers, I think teams would be constantly demanding packages built around hitting when dealing a valuable MLB asset. I just haven’t seen that happening.
Me too, soon
(I’m still working through the comments on the Yankees Top 20 before I call it a night). I’m not saying that the difference is as stark, though, and I think I’ve been clear on that. I would except a difference, but a smaller one. Ad I think adding more data, if it’s available, would go a long way.
i'm totally confused by what you are trying to argue
put yourself in Billy Beane’s shoes
you have Gio Gonzalez and a $40 million budget to work with
you know when you trade Gio Gonzalez you will get, say, $30 million in surplus value back in the form of prospects, regardless of whether you trade for pitchers or hitters. do you choose to trade for pitchers or hitters?
if i were Billy Beane, i would always trade for pitchers. this is because i want my most expensive assets to be my least volatile. thus, i’d want my pitching prospects (most volatile) to also be my cheapest. and i’d want my hitters that i get through free agency (much more expensive assets than pitching prospects) to be a lot less volatile.
how does this not make any sense?
and what does Wang’s research have anything to do with it?
We've been discussing Wang's research?
You do make an interesting point. My counter would be that it would be hard to get as much surplus value if you’re dealing solely for less valuable assets. You don’t think anyone would offer comparably rated hitters that Beane should jump on because it would hold a lot more surplus value?
And why did he trade the one really valuable hitting prospect he’s had recently without getting nearly as much surplus value as Wang’s research suggested he would be worth.
i'm not sure what you mean by
getting less surplus value for less valuable assets, unless that was a typo and you meant “volatile”
whether Gio Gonzalez is traded for hitting prospects or pitching prospects, the expected value should be the same.
what i mean is, Billy Beane is only going to trade Gio Gonzalez, if he expects the surplus value he gains exceeds the surplus value he loses. he should, hypothetically, always simply accept the trade package that he thinks offers the highest surplus value, with some caveat to the volatility aspect as i mentioned above (due to his exceedingly low budget, he should always be minimizing volatility in expensive assets and maximizing volatility in cheap assets). through some combination of these two factors, Beane’s traded Cahill and Gio for packages largely centered around pitching prospects.
as for your question, if i were Beane, and i received two trade offers for Gio Gonzalez, and i fully believed each held the same surplus value, then i would always pick the pitching package over the hitting package
i honestly don’t know which hitting prospect trade you are referring to. are you referring to Carlos Gonzalez?
i think if you take a look at the CarGo trade, you will see it fits my framework perfectly. Beane traded CarGo and Street for Holliday. CarGo was a projected $22 million in surplus value based on Wang’s research, and Street was worth $2 million in surplus value in 2008. expected arb raise would shave that expected surplus value down by $1 million. Holliday was worth $19 million in surplus value in 2009. expected arb raise would shave that down to $15 million.
so Beane traded away $23 million in surplus value in order to get $15 million in surplus value, which is what you would probably expect if 1) exchanges occur when one party undervalues their own assets and overvalues the counterparty’s assets; 2) trades are auctions where the highest bidder wins, and thus overpay above the expected value occurs; and 3) you pay a premium in order to obtain less volatile assets. major league hitter is less volatile than minor league hitter/relief pitcher.
as you can see, the CarGo trade is something Beane would definitely do if he was operating under the framework of volatility minimizing i outlined above, and if he believed in the surplus value chart above.
Beane
He hasn’t had a winning season in 5, going on 6 years, so perhaps he should rethink the strategy:)
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Or
He shouldn’t have told the world his strategies in a book. =P
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
I think
There’s a valid point here. Beane seized on a market inefficiency that for the most part is gone. It doesn;t appear like he’s been able to find new market inefficiencies, the way the Rays did.
The new CBA dropped a new inefficiency right in Beane's lap
and the laps of about a dozen other guys.
Now we’ll really see how smart he is.
The point is he's aware of it
And he clearly ignores it despite the fact he is a guy who searches for undervalued assets constantly.
No one is saying the numbers are exact and to be taken as gospel. Multiple posters have acknowledged that the numbers may be off, for various reasons.
This goes beyond just being inexact
He has literally done the exact opposite of what Wang’s numbers suggest is the “right” strategy.
So one GM is going against what some research suggests, a GM who hasn’t exactly done real well the last couple years, and that means the research is invalid.
I want to bring in a piece of stock market wisdom here, I think it’s apt. Each trade has two sides – if you’re on one end, there’s somebody on the other end, taking the opposite view as you are. So you better know who the idiot in the room is, or chances are it’s you, and you’re going to get whooped.
Well, likewise, there’s other teams on the other sides of BB’s trades – so I wouldn’t be so sure the BB is fleecing all of them over the place, maybe he’s the one getting the short stick if he’s insisting on pitchers (we have no idea of this)
You're missing my point
The guy who really helped start the current trend to a more mathematical look at baseball is doing the exact opposite of what some basic mathematical research is telling him about value. Either he somehow missed this piece that tons of random internet posters are aware of, or he’s actually had his people crunch the larger amount of data he has access to and the results are what is currently guiding him.
I have no doubt
And I’ve already said multiple times, that I’m sure the numbrs are imperfect, and for pitching prospects, highly imperfect. Just just because one guy is doing one thing, does mean you ca throw out the research, and more importantly, the framework (which can be improved on with more data). If the guy were very successful, then sure, maybe you would, or at least b highly skeptical. But it’s not like Beane is picking everyone’s pockets when he makes these trades the last couple years.
Lets get to the basics...
Do you think Beane, with multiple people working for him whose only job is to do things like this, has access to more extensive studies than what Wang put out there for free on the internet largely on his own time? If so, shouldn’t Beane’s actions inform us more than Wang’s research?
And as I said before, my issue isn’t with the framework itself. Its with the time lag and the fact that without many more years of data, we are left with an incomplete picture of which prospects were actually producing the most value (for example, how much more valuable would pitching prospects look if we could look at guys who came up in the 80s).
To your first paragraph, again, you neglect that there are two sides to each trade, and each of these other GMs has a similar number of peple working for him pumping out similar data. So if they’re doing the opposite of Beane, maybe it’s Beane who had the inferior info. It’s not just beane’s action vs. Wang’s research, it’s Beane’s actions vis-a-vis his fellow GMs that we should analyze. Is he the genius in the room, and everyone else the idiots (with his info opposite in conclusion to Wang’s) or is he the one being taken advantage of? This is not at all clear. You are ignoring what all the other GMs may or may not know.
To your second paragraph, I;m agreeing! Let’s add as much data as possible. Including all the pitchers (and hitters) from the 80s. Of course this will make the results better and more useful.
Look at the teams he traded with
Mid 2000 Cardinals (not know for progressive research) and Braves (last I heard they employed one stat guy in the entire operations department and that was after the trade). The Dbacks aren’t exaxtly a team known for being on the cutting edge of research either. And remember we talking about times where Beane got to pick from multiple offers. It doesn’t really matter if one team is smarter than him if others that made an offer aren’t.
Another thing
For the most part, Beane is getting multiple pitchers (at least it seems that way). I’d really like to see full data on this. A list of every player traded before they exhausted there prospect status shouldn’t be too hard to come up with.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Other than the Gonzalez deal, not really
Hudson, Mulder, and Cahill were all built largely around one really good pitching prospect
Yes and no
Even with the Haren-to-AZ deal, Brett Anderson seemed to be a big part of the deal. 1a and 1b perhaps.
by nobodyinparticular on Jan 2, 2012 3:22 AM EST up reply actions
Injuries
Billy Beane in moneyball (if half the stuff is more factual than fiction, which i know the movie was a stretch but some of it derives from actual events)
Knows what he’s doing, He has the right mindset and you know should have been Boston’s GM before Epstein was even offered. (where would Theo be today? had Beane taken the Sox job back in ’02-or ’03 or whenever it was exactly)
Yoenis Cespedes
http://www.youtube.com/watch?v=aW9ge8l3jY8
by SteveHoffmanSlowey on Jan 2, 2012 7:14 PM EST up reply actions
verlander wouldnt be involved...but neither would adam miller...
it evens out
Fire Everyone
by billybeingbilly on Jan 2, 2012 12:09 AM EST up reply actions
You're missing the point
I’m not saying Verlander would somehow change the numbers by himself. I’m saying the fact that a guy who has established himself as one of the top pitchers in the game over the past 5 years isn’t part of the current data set points to the fact that the data isn’t current enough.
And this isn't a criticism of Wang's work
Its not really a criticism at all. Its just an issue with trying to put a value on prospect rankings. You have to wait until so long after the players are prospects to evaluate how good they actually were, that the data just isn’t that relevant any longer.
Improvement
First, I would only count each player once. Whether you go highest ranking or ranking at the time he broke into the big leagues is a tough choice. Maybe I’d do it both ways, then also for the first X years after his last ranking, regardless of MLB service time. Or possibly the guy could be in each bucket a max of once. One thing you don’t want is a Chipper Jones, who was a top 10 prospect four times, counting quadruple a guy who was only a top 10 once. In fact, I can’t believe Wang did that. IMO, this is the biggest problem with the study and in fact would render it moot for the top 10.
Second, I’d use straight rWAR. I definitely wouldn’t translate to $. Whether to discount. or not is an open question. I’ve done it both ways and I prefer not discounting.
Third, I’d try and use Sickles grades as much as possible. Some years there are better prospects than others, especially at the high end, which is where the real value (and only discrepancy) is. Sometimes a top 10 prospect might be a straight A, sometimes he might only be an A-. IMO, there is a big difference between the groups. Sickles started in 1995, so where getting to the point where he’ll have a decent sample size.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
I didn't realize he started that early
It would actually be viable to do that kind of study…in theory, without knowing if the grades back then were similar to now in terms of A, A-, etc, you could do 10 years 1995-2004 and get decent numbers. The question is, is this even available, since I assume most of those books aren’t easily available.
In regards to what you suggested…I’d use a 50/50 split of rWAR and fWAR, though it’s more work. But I agree on using WAR, and not $. I would think discounting makes sense, since there is a time value.
Finally, the counting issue is the hardest one. I can see valid arguments both ways. As you say, it introduces a bias when you have the same guy 4 times. On the other hand, it’s a hell of a judgement call otherwise.
Actually
If you were doing it according to grades, what I think I’d do is include a player in each grade category in which he was ranked, but just once so as to not bias the data. So let’s say a high schooler player was ranked: B+, A-, A, A…I’d include him in three buckets, but not double count the A. In this sense, we’re answering the question: A player receiving grade X will on average produce X WAR.
This way, each grade bucket would include players who may have later moved higher or lower, but were ranked in that bucket. But it wouldn’t double count someone like Jones or Bryce Harper, who would both have been pure A on multiple occasions.
Does this make sense?
Right
That’s probably the best idea.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
And you could do the same thing for the BA
Not double count them in buckets…only count a top 10 prospect once, regardless of how many times. Of course, you end up with trouble like what is a guy is 15 one year and 9 the next…the judgment may not be different, but the bucket would be.
Sorry to kinda spam
But you’ve kinda got me hooked on the idea, and I’m thinking new angles every time. The other great thing about using grades is, there’s not a finite number of each…there have to be 10 top 10 players each year, but not ten grade As, etc.
Exactly
With the Bs and Cs, the sample is big enough and the dropoff is small enough that it won’t matter. It would be huge for the As and the A-s.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Time Value
The time value is for money. I’m not sure how much it goes for WAR. If you want to buy 1 WAR today, it will cost you $5m, but next year it might cost $5.5. That’s why the translation to $ is silly and why I prefer not to discount WAR.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
True
Right, I kinda forgot about that. I need to think it through a little more, but what you’re saying makes sense. I think (again, this is preliminary) it depends on what you’re trying to answer as well. If it’s just production, or if it’s value to the team.
WAR today is worth more than WAR in the future
The thinking behind it is that more WAR leads to more wins leads to more money leads to the ability to buy more WAR
fWAR
I don’t like to use fWAR when looking backwards, especially over a 6 year period. With such a huge sample, the extra work wouldn’t add anything extra information and might in fact taint the sample for pitchers.
I thought John had PDFs for sale of all the old books, but that would be pretty expensive. He’s also done reviews of his 50/50 lists for virtually every year. I can’t recall if those have the letters though.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
I think you’re right, just on a strict cost/benefit analysis. It’s a lot more time, and in aggregate, won’t make a difference at all.
Another Thought
I’ve generally used the first six years as the way to rate how the prospects performed, but I’m starting to flip-flop on that. I think there is a big problem comparing some guys age 20-25 seasons with others age 23-28. I also don’t think scouts are projecting what they can do over the next six years and I don’t think John is grading on that. They are projecting/grading what that player could be at his peak.
The other issue is, it is the teams that control when a prospect gets called up. A team like the Rays gets rated highly for prospect development when in reality all they are doing is leaving them in the minors longer. Yeah, that gives the team more value, but does it change how good the player eventually becomes?
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Very good point
Something I was thinking about as well. I guess the age thing is priced in to a certain extent by discounting WAR to present value (players do get some extra credit for getting called up quicker), but that has to do with the teams themselves, not the underlying value of the prospects.
And it does seem like a bit of an issue that prospect rankings aren’t trying to project value in the same way that Wang’s system (not to further that argument, just as an example) does. There rankings would probably look different if they were purely trying to project value for the first six seasons.
I agree with where you're coming from
But ultimately, I think it comes down to what you’re trying to.
Wang and subsequent research were trying to estimate the value to the team, which is essentially the surplus value they get over market, which is really just the pre-free agent years. So it makes sense to me to only look at that period (and to be transparent, this is what I’m really interested in).
On the other hand, you can try to evaluate how good the player actually turns out to be, which I;m guessing is what you’re more interested in. For this type of exercise, I completely agree you’d want to not limit yourself to 6 years.
Perhaps
They take the pitchers because they can’t get the hitters.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
You really think no one offers hitters?
I mean pretty much every major trade Beane has made has been built around pitching. Gonzalez, Cahill, Mulder, and Hudson were all traded for packages built around pitching. You really think no one offered him hitting for any of those guys? And even if you thought that was true, why would Beane trade Carlos Gonzalez a year after he was ranked the #22 prospect in baseball? What’s more likely, Beane hasn’t seen Wang’s research, or Beane sees the same flaws in it that I’m pointing out?
it would be ambitious...
…but you could actually do some research to see win shares or some value based on previous years’ grades and assign values based on real value rather than on nominal values – tho, truth be told, I wouldn’t put that effort in either. I would change the values however to create more separation within the rankings/values…
-peter
by PeterF on Jan 1, 2012 6:43 PM EST via iPhone app reply actions
but you could actually do some research to see win shares or some value based on previous years’ grades and assign values based on real value rather than on nominal values
The issue is that I don’t think enough years have passed by to determine the real value of the grades. I liked the weighing system done by Doug.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Its Victor Wang's research, Doug just uses it
And again, its an extremely flawed system that doesn’t tell us much other than who has the most good hitting prospects.
Degree
Saying the system is extremely flawed seems a bit extreme. I think we can all agree that the grade A pitchers should be worth a little more (probably about 20), but the rest of that chart and the theory behind it is very sound. Outside of the top 10, it’s not like it’s a small sample that can be skewed by 1-2 pitchers.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Outside of the top 10, it’s not like it’s a small sample that can be skewed by 1-2 pitchers
Well first of all it was more than 1 or 2, and even if you want to look at the top 25, that’s still only a sample size of 102 pitching prospects (double counting guys). Just looking at 3 major busts (Todd Van Poppel, Brien Taylor, and Roger Salkeld), they account for 9% of the the pitching list. Besides, like I said, the research is based on a time period where teams simply weren’t producing good pitching prospects, which severely skews the stats. Just look at the pitching talent that came up in the 80s and 00s, and you’ll see how the relative dearth of talent in the 90s kind of messes the whole thing up.
Impact
I think you’re overestimating the impact a few guys can have. We’re only talking about the first six years of a players career, so the top end is going to be about 30 WAR. When you’re talking about a sample of over 200 pitchers (anything outside the top 10), a couple of guys aren’t going to make a huge difference.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
30 WAR is most certainly not the top end
Clemens was just short of 40 in his first 6 years. Let’s say he was a prospect in the 90s and made the top 10 once. By himself he would have upped the average for the group by over a win. Just as a rough estimate (I know Wang actually discounted to present value), Clemens alone would add 4 or 5 million in average value to those top 10 prospects. And that’s if he was ranked 1 time. If he made it 2 or 3 times like some of the busts, you’re talking about one player adding anywhere from 8-15 million to the average value of top 10 pitching prospects.
Now imagine if we weren’t just talking about Clemens, but also Maddux (29.7 WAR), Smoltz (21.5), etc.
Degree
So again, we’re on the same page on nearly everything. The question is, to what degree is the top 10 tainted?
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
To the point where its worthless with regards to elite pitching prospects
I mean according to his research, top 10 pitching prospects are less valuable than lower rated guys. If that doesn’t tell you the numbers can’t be completely trusted, I don’t know what will.
Or it tells us
That pitching prospects are infinitely harder to identify as elite than hitting prospects are. And that still holds its value in recent times. We all debated who was the #1 pitching prospect in baseball between Homer Bailey and Phil Hughes. Neither guy has even come close to justifying that ranking to this point in their careers (I still like both guys in the future).
For every Homer Bailey and Phil Hughes
There is a Delmon Young and Alex Gordon.
The only basis of an argument I will buy that hitting prospects > pitching prospects is one that centers around susceptibility to injury. I do not believe in TINSTAPP at all.
Alex Gordon and Delmon Young
Alex Gordon has more than TWICE the WAR of Bailey and Hughes combined.
Delmon Young has more WAR than both of those pitchers combined.
I don’t believe in tinstapp either. In fact I get in arguments often about it. What I do believe in is that for the most part, we (scouts, writers, teams) have troubles correctly identifying the best among the top 50-100 pitching prospects in the game a lot more than we do with the same number of hitting prospects.
I think it is fair to say though
That both Gordon and Young did not become the players both we and scouts thought they would be – despite the height Gordon reached this past year.
We can agree to disagree here. I don’t think it is fair to say scouts and analysts have an easier time rating hitters than pitchers. A big majority of pitchers flame out due to injury – not necessarily skill advancement (or lack thereof). If anything, as an armchair analyst I find it easier to rate and rank pitchers.
I think we are in agreement
But the fact that pitchers flame out so much due to injury is kind of the point isn’t it?
Maddux
I know I’m being kind of nitty here, but there is no freaking way Greg Maddux was ever a top 10 overall prospect. His prospect peak would have been 1986, which was a huge year (Canseco, Clark, Palmiero, Surhoff, Larkin, Santiago, and Swindell off the top of my head). Looking at Maddux’s numbers, where he was drafted, and what we know about his stuff, I’d guess he was a B, maybe a B+ at best.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Are we grading stuff based on FB velocity?
Maddux could get insane movement on both his fastball and his change towards whichever side of the plate he wanted. This idea that he didn’t have great stuff is just funny (and for the record, he sat comfortably in the low 90s early on in his career). Throw in good control at the time, great performances in terms of ERA (and let’s not pretend that wasn’t incredibly important in rankings then), and the ridiculous inning loads he showed the ability to handle and I think he would rank quite highly. I mean post1986 he was coming off a sub 3 ERA primarily in AAA over 192 innings (over 7 1/3 per start) as a 20 year old. Sounds an awful lot like Teheran to me.
Maddux Retro
http://www.minorleagueball.com/2005/8/23/14716/4363
The difference between Maddux and Tehran is, Tehran was an A coming into his age 20 year. Maddux was a B-. That’s going to put Maddux at B+ at best, most likely a B.
Looking back at the top pitching prospects heading into 1987, the guys who were likely better than B+;
Greg Swindell-Definitely rated higher than Maddux. He was likely the top pitching prospect in the game.
Joe Magrane-Also very likely a straight A grade.
Eric Bell, Willie Fraser, and Mike Campbell-These guys look like A- types and very likely rated higher than Maddux on prospect lists.
Here’s the guys who look like B+ prospects; Todd Stottlemyer, Tom Gordon, Bill Wilkinson, Rich Sauveur, Jim Medinger, Steve Searcy, John Burkett, Mike Cook, Tom Glavine and possibly Melido Perez.
Also you had Kevin Brown and Kent Mercker selected in the top 5 of the 1986 draft that could have been looked at as B+ material.
Other guys who were prospects at that time; David Cone, Jeff Brantley, Chris Bosio, Shawn Hillegas, Randy Myers, Mike Dunne, Duane Ward, Rob Dibble, Ken Hill, Tommy Greene, Darren Holmes, Juan Guzman, John Smoltz, Mike Scott, Kevin Tapani, Ramon Martinez, and John Wetteland. Most of those guys were C+ types with sleeper upside.
Once you factor in the hitting prospects, there is just no way that Maddux was in the top 10 overall.
Also, looking back, my guess would be that John Smoltz never made the top 100 prospects. He was really just a stuff guy with no performance until 1988. Problem is, he exhausted his prospect eligibility before he could be ranked.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Just as example, how much higher would the 80s values for pitchers look?
You’d likely have guys like Clemens, Maddux, Smoltz, Schilling, etc. bringing those values WAY up.
Selective Sampling
You are just selecting the best pitchers (for their career) instead of looking at who were the top prospects. The only one of those guys who was likely in the top 10 overall was Clemens (he was also the top producer). Smoltz and Schilling might not even have been top 100 guys!
And how do you not even mention the top pitching prospect of the 80s-The Original Strasburg, if you will-Dwight Gooden. He was also a complete stud over his first six years, racking up 30 WAR.
Here’s the list of hitters that debuted during the 80’s that put up more WAR than Clemens over their first six years; Boggs, Bonds, Griffey, Ripken. Griffey and Ripken were both #1 overall type prospects and Bonds was definitely a top 10 guy.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
No I looked at guys who had the sort of stuff I thought would get them ranked highly
And I’m not saying top 10 pitchers are more valuable that top 10 hitters, so comparing those guys to top hitters is meaningless. I’m saying that top 10 pitchers are clearly worth more than 26-50 hitters. Unless you think teams are completely ignorant of Wang’s research, then you have to admit that clearly they are. I can’t believe people really want to argue this. Since when is 10 years of prospect data enough to draw the sort of wide-ranging conclusions that people draw from Wang’s research?
I agree the conclusions drawn are two wide ranging, and used too rigidly to analyze, rather than using them as a nice starting point which is how I see them as most useful.
It’s not that teams are ignorant of the data, they just have a lot more information, period. Whereas BA is one source, they can compile the data gleaned from scores of scouts over scores of years. And I’m sure they use this type of analysis to inform the way they think about trades.
Obviously they have more data
And pitching doesn’t trade at a significant discount to hitting, which I think kind of shows that Wang’s data really isn’t all that informative when it comes to pitching prospect value.
Data
Where is the data to back this up? Because to me, it sure seems like top pitching prospects get dealt more often than top hitting prospects.
Fight for licensed online poker in 30 seconds. Take part in the daily action plan!
https://www.facebook.com/pages/Poker-Players-Alliance-Daily-Action-Plan/240644152648049
twitter @PPADailyAction
Perhaps they get dealt more
Because the team trading the veteran prefers to get pitching. You’re acting like the team dealing the prospects has the all the power, but the team dealing for the prospects gets to listen to every offer and then select what they feel is the best one. I’m just not buying that no team is willing to deal a good hitting prospect.
Think about the Beane example I gave you. You really don’t think he got offered any good hitting prospects for Mulder, Hudson, Cahill, and Gonzalez? Because he dealt all those guys mostly for pitching. And then he actually traded CarGo 1 year after he ranked #22 in the game per BA, so it sure doesn’t seem like the guy who you would think would be on the cutting edge of this stuff agrees with the idea that hitting prospects are significantly more valuable than pitching prospects.
With Doug's work
I feel like I would just be copying his work. As such, I won’t update this any longer unless certain aspects of the weighing system is updated.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
I like your analysis better quite honestly
I fully endorse the KISS method.
If there was one thing I would change to your grading system, it would be to somehow weigh A and A- grades more heavily (and maybe B+ as well). Maybe start with a 5.00 grade for an A+ and work your way down to a 3.00 for a B.
Maybe I should have kept it simple and basically said this
I think the sort of system you’re talking about is pretty much just as enlightening, while not really giving the false since of accuracy that Wang’s system does.
Good
stuff. Pirates are higher than I expected. Guess an A and A- go along way.
by McCutchenIsTheTruth on Jan 1, 2012 7:23 PM EST reply actions
Oakland A's
surprised to see them not above the last 5 teams on this list after the trades they made.
John hasn't done their list yet
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Neat stuff.
I’m sure many of the concerns about the absolutely use of this ranking method are valid, but it’s a really cool idea nonetheless.
"I'll be deep in the cold, cold ground before I recognize Missoura!"
Agreed
Anything that can attempt to weigh a larger list of prospects and give us team rankings is a good thing. I think that often the “national rankings” are mostly just the top 10 guys from each team and farm systems are a lot more than that.
I think the Pads and Jays are closer at #1 and #2 then it appears
really SImon Castro is a White Sock now and The Jays system has more depth than just 20-24,25 guys than anybody else in the top tier IMO.
+ personally I"d Without HESITATION swap the Mariners and the Red Sox, sure the red sox have a lot of depth…
But the M’s have some realy impact guys the Sox simply don’t have… i could see from both their respective pools of top prospects 4 years from now the Sox develop a decent OF with a bunch of 4th OF / Util types + a decent but non-star 3B and Catcher and maybe a middle infielder who can defend with limited pitching.
Where as the M’s could have ONE HECK of a Starting Rotation (even without King Felix) in just a few seasons.
I’m suprised the most by Bo Sox over Mariners.
Yoenis Cespedes
http://www.youtube.com/watch?v=aW9ge8l3jY8
by SteveHoffmanSlowey on Jan 2, 2012 7:09 PM EST reply actions
+ personally I"d Without HESITATION swap the Mariners and the Red Sox, sure the red sox have a lot of depth…
But the M’s have some realy impact guys the Sox simply don’t have… i could see from both their respective pools of top prospects 4 years from now the Sox develop a decent OF with a bunch of 4th OF / Util types + a decent but non-star 3B and Catcher and maybe a middle infielder who can defend with limited pitching.
The new grading system brought the two a lot closer than they were previously:
Original:
9. Red Sox – 2.7255
10. Mariners – 2.6670
New:
9. Red Sox – 2.7510
10. Mariners – 2.7380
That’s a decrease in net difference from 0.0585 to 0.013.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Impact Prospect Debate
I certainly see the three/four impact guys in SEA’s system (Hultzen, Walker, Paxton, Campos), but do you see any impact bats? I think you’re selling the Sox system short . As with all spects, if’s are attached to them, but certainly Xander, Swihart, and Barnes should fit the “impact” criteria.
ProspectTube.com
You Video. You Scout.
by ProspectTube.com on Jan 3, 2012 10:46 AM EST up reply actions
Update
With the recent A’s list, I am done with the first part. I’ll update once John has the final grades posted.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Looks good
I like how the results ended up turning out
And John appears to be done
So this will likely be the final rankings
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Minor League Ball
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
