Braves rated the best farm system of the decade by MLB

Metaphysicist said:
Here are the whole draft #s, 1-40/50; I was right, it makes Wren look worse.

2005-2009:

$226M Expected Surplus Value
$803M Actual Surplus Value (I think I miscounted before, actually)

+332M over-expectation outside top guy ($247M from Freeman)

Freeman: +247M
Heyward: +199
Escobar: + 101
Kimbrel: +99
Medlen: +76
Hanson: +63
Minor: +26
Hale: +9

2010-2014:

$158M Expected Surplus Value
$234 Actual Surplus Value

-73M shortfall outside of top guy ($153M from Andrelton)

Simba: +153
Wood: +62
Gattis: +30
Webb: +7
Gosselin: +6
S. Simmons: +4
Shreve: +3
Cunningham: +1
Click to expand...

I'm wondering about some of these numbers.

Wood generated 13.6 fWAR in his pre-free agency years. Kimbrel 13.5. Yet you have Kimbrel at +99 and Wood +66.

Simmons generated 23.2 in what would have been his pre-free agency years. Heyward 25.1. But you have Heyward at +199 vs Simba at +153, which seems to overstate the difference in WAR by quite a bit.

Freeman comes in at 20.9 WAR in his pre-free agency period. Yet you give him a whopping 247M surplus value.
 
Last edited:
(a) These are numbers only as Braves. Except for a few specific situations, I don't think it's appropriate to credit what someone does multiple years after leaving the organization to Braves' drafting and development. That's adding a lot of noise to the data. You could maybe add in trade receipts, but that's adding a whole separate form of noise.
(b) Using bb-ref WAR. Using FIP for pitchers retrospectively is nonsensical.
(c) Freddie has never hit free agency; his current extension is part of the value of drafting him.
(d) Doing this probably doesn't map exactly because of the changing $/WAR.
(e) These #s are all very back of the napkin. I wouldn't take them super seriously.

But if you prefer, any team, no extensions. But still using real WAR:

2005-2009:

$226M Expected Surplus Value
$969M Actual Surplus Value

Heyward: +250
Escobar: +195
Freeman: +175
Kimbrel: +99
Medlen: +67
Hanson: + 55
Flowers: +39
Minor: +26
Oberholtzer: +24
Devine: +17
Hale: +14
Hoober: +4
Rasmus: +3
Hicks: +1

2010-2014:

$158M Expected Surplus Value
$557M Actual Surplus Value

Simba: +284
Wood: +89
Ahmed: +86 (feels ridiculous to "credit" him to us, but so be it)
Gattis: +58
LaStella: +20
Shreve: +17
Caratini: +8
Gilmartin: +4
Gosselin: +3
Drury: +3
S. Simmons: +3
Cunningham: +1


So Clark:

$743M over expectation
428% of expected value
317% (outside Heyward)

Wren:

$399M over expectation.
353% of expected value
177% (outside Simmons)
 
I'll just note that usually when looking at expected value of a draft pick or prospect, the valuation models look at value generated during the 6 (plus a fraction in some cases) pre-free agency years. You don't count team friendly extensions beyond that period and don't allow the fact that the draft pick or prospect was traded before hitting free agency to make a difference.

So to determine value under those constraints you would look at Freeman WAR those pre-free agency years, and ditto for Heyward, Simba and others.

The point of the exercise we are doing in this thread is to evaluate drafting astuteness. And I think constraining the valuation model with the conditions listed above allows us to get as close to a fair answer as possible.
 
Last edited:
Metaphysicist said:
(a) These are numbers only as Braves. Except for a few specific situations, I don't think it's appropriate to credit what someone does multiple years after leaving the organization to Braves' drafting and development. That's adding a lot of noise to the data. You could maybe add in trade receipts, but that's adding a whole separate form of noise.
(b) Using bb-ref WAR. Using FIP for pitchers retrospectively is nonsensical.
(c) Freddie has never hit free agency; his current extension is part of the value of drafting him.
(d) Doing this probably doesn't map exactly because of the changing $/WAR.
(e) These #s are all very back of the napkin. I wouldn't take them super seriously.

But if you prefer, any team, no extensions. But still using real WAR:

2005-2009:

$226M Expected Surplus Value
$969M Actual Surplus Value

Heyward: +250
Escobar: +195
Freeman: +175
Kimbrel: +99
Medlen: +67
Hanson: + 55
Flowers: +39
Minor: +26
Oberholtzer: +24
Devine: +17
Hale: +14
Hoober: +4
Rasmus: +3
Hicks: +1

2010-2014:

$158M Expected Surplus Value
$557M Actual Surplus Value

Simba: +284
Wood: +89
Ahmed: +86 (feels ridiculous to "credit" him to us, but so be it)
Gattis: +58
LaStella: +20
Shreve: +17
Caratini: +8
Gilmartin: +4
Gosselin: +3
Drury: +3
S. Simmons: +3
Cunningham: +1


So Clark:

$743M over expectation
428% of expected value
317% (outside Heyward)

Wren:

$399M over expectation.
353% of expected value
177% (outside Simmons)
Click to expand...

Looking at fWAR I have Yunel at 19.3 over the pre-free agency years.

So by fWAR Yunel<Freeman<Simba<Heyward

By your numbers Freeman<Yunel<Heyward<Simba

But I think this iteration is a pretty good summary of the two eras. Thanks for crunching all the numbers. I think it provides a couple conclusions:

1) The 5 years prior to the Golden Wren Era produced a better yield controlling for the expected value of the picks.

2) Both eras actually did good drafting work relative to value of picks. Wren got canned in part because of a couple bad drafts, but his overall record is not bad.

One last point. Some of the Golden Wren Era picks still have some service time left on their clocks. Ahmed has one season left. LaStella has a year left. Guys like Webb, Sobotka, Caratini, Drury and Sims have multiple years left. So the books on the Golden Wren Era drafts are not quite closed yet.

And the books on the Hartcoppy drafts (2015-2017) remain very much open.

One last last point (said in my best Lieutenant Colombo imitation). The numbers from the Roy Clark group add up to 933. And from the Golden Wren Era drafts to 576. Doesn't change the basic conclusions.
 
Last edited:
Yeah Wren got canned cause we under performed in 2014, he locked us into a few really terrible contracts, and he cleaned out the farm with no immediate help in sight due to poor draft classes. Sort of a prefect storm of scenarios.
 
Carp said:
Yeah Wren got canned cause we under performed in 2014, he locked us into a few really terrible contracts, and he cleaned out the farm with no immediate help in sight due to poor draft classes. Sort of a prefect storm of scenarios.
Click to expand...

Now if only the same set of conditions could have gotten Suspenders fired after the Teixeira and Millwood deals...
 
Snowman said:
Now if only the same set of conditions could have gotten Suspenders fired after the Teixeira and Millwood deals...
Click to expand...


Edit: Disregard. I thought Tex played for LAA for another year after the trade.

Still, I am not sure which trade was worse, the original Tex trade or the sequel. We sure gave up a King's ransom for him, but with the 2 picks the Angels received for letting Tex walk, they drafted Mike Trout and Tyler Skaggs. We got punched in the nuts twice with the Tex trades.
 
Last edited:
Carp said:
Edit: Disregard. I thought Tex played for LAA for another year after the trade.

Still, I am not sure which trade was worse, the original Tex trade or the sequel. We sure gave up a King's ransom for him, but with the 2 picks the Angels received for letting Tex walk, they drafted Mike Trout and Tyler Skaggs. We got punched in the nuts twice with the Tex trades.
Click to expand...

It's definitely the first one. A lot of teams, including the Braves, passed on Trout. Doubtful the Braves pick him with that pick anyways. Both were horrible trades though.

Sometimes I'd like to imagine what adding Elvis, Feliz, and Harrison to those 2010-2012 teams would of done.
 
thewupk said:
It's definitely the first one. A lot of teams, including the Braves, passed on Trout. Doubtful the Braves pick him with that pick anyways. Both were horrible trades though.

Sometimes I'd like to imagine what adding Elvis, Feliz, and Harrison to those 2010-2012 teams would of done.
Click to expand...

Amazing how much smarter we are with 7 years of hindsight.

Not sure adding them would of done very much - it might "have" done something though.
 
All but the dumbest IBWAA members knew the Tex trade was bad the moment it happened. No hindsight was required.
 
nsacpi said:
Looking at fWAR I have Yunel at 19.3 over the pre-free agency years.

So by fWAR Yunel<Freeman<Simba<Heyward

By your numbers Freeman<Yunel<Heyward<Simba

But I think this iteration is a pretty good summary of the two eras. Thanks for crunching all the numbers. I think it provides a couple conclusions:

1) The 5 years prior to the Golden Wren Era produced a better yield controlling for the expected value of the picks.

2) Both eras actually did good drafting work relative to value of picks. Wren got canned in part because of a couple bad drafts, but his overall record is not bad.

One last point. Some of the Golden Wren Era picks still have some service time left on their clocks. Ahmed has one season left. LaStella has a year left. Guys like Webb, Sobotka, Caratini, Drury and Sims have multiple years left. So the books on the Golden Wren Era drafts are not quite closed yet.

And the books on the Hartcoppy drafts (2015-2017) remain very much open.

One last last point (said in my best Lieutenant Colombo imitation). The numbers from the Roy Clark group add up to 933. And from the Golden Wren Era drafts to 576. Doesn't change the basic conclusions.
Click to expand...

I'm not using FWAR (well, maybe I should for catchers?). For pitchers it uses FIP retrospectively, which is stupid. For defense, it uses UZR which is worse than DRS (last I read, at least). All the numbers are based on bWAR.

On Yunel:

(a) I have him at 23.2 bWAR through 2013, and bbref has him with 5.121 service time entering that year.
(b) DRS likes him better than UZR, so that's a couplafew of wins right there. Same, but more so, for Heyward, and especially for Simba.
(c) He also comes out better because he got paid the least. Pre-FA:

- Yunel: ~14M - 23 bWAR
- Freeman: ~27M - 23 bWAR
- Heyward: ~17M - 30 bWAR
- Simba: ~$29M - 35 bWAR

Example of what I'm doing from 2005:

Hmhk8ly.png


Anyway, I'd agree with both your conclusions. Wren's drafts were not as good as Clark's, but were overall very positive. Clark's numbers are probably even exaggerated a little because WAR theoretically gets cheaper as you go farther back in time. If I were gonna do this again, I would have figured out where FG was generating their $ numbers and just used surplus WAR directly.

But frankly, given all that, I still wouldn't have been 100% confident about his drafts going forward. Look at the shape of his success; there is a striking downslope in quality:

2010 was fabulous.
2011 looks good retrospectively, now that 2 backup IFs, Ahmed and LaStella, have blossomed 5 years after we gave up on them. They've saved what looked like a complete wash.
2012 is a winner thanks to Wood
2013 straight up bad, well below expected value thus far (though maybe Caratini becomes a star)
2014 Also currently below expected value, unless Webb/Sobotka/Davidson saves it

Kw4woZa.png
 
nsacpi said:
I'll just note that usually when looking at expected value of a draft pick or prospect, the valuation models look at value generated during the 6 (plus a fraction in some cases) pre-free agency years.
Click to expand...

I'm fine with whatever cutoff you wanna use, and that one makes sense to me; but just to be clear, the Fangraphs article you posted, and from which I was pulling the numbers, was not using "first 6 MLB seasons" to generate the data, but first 10 seasons after being drafted for college players and first 11 seasons for high school players. Thus the extra years for Freeman.

nsacpi said:
You don't count team friendly extensions beyond that period and don't allow the fact that the draft pick or prospect was traded before hitting free agency to make a difference.

So to determine value under those constraints you would look at Freeman WAR those pre-free agency years, and ditto for Heyward, Simba and others.

The point of the exercise we are doing in this thread is to evaluate drafting astuteness. And I think constraining the valuation model with the conditions listed above allows us to get as close to a fair answer as possible.
Click to expand...

That's fair. I mean, at least as far as guys like Heyward and Simba are concerned. They were essentially the same players both before and after a trade. The main purpose of the constraints I added was to deal with guys like Ahmed who we didn't seem to think much of, traded them as prospects, and who then spent 5 years being trained by another organization before getting good. I think "counting" him is misleading as to our draft astuteness; I don't think you can assume he does the same if he stays a Brave (unlike Heyward's final pre-FA season).
 
Metaphysicist said:
If I were gonna do this again, I would have figured out where FG was generating their $ numbers and just used surplus WAR directly.
Click to expand...

Yeah, I find it cleaner to work with WAR directly. Also a couple useful shortcuts. Surplus value for the pre-arb years is pretty much 100% of WAR and for the arb years is 50% of WAR. Additional shortcut from combining those two, surplus value is about 75% of WAR from the pre-free agency eligibility period. These are obviously approximations but if applied consistently across samples will give you a pretty accurate result for the kind of discussion we have been having.
 
That's probably fine for looking at big datasets, but not sure how accurate those percentages really would be on a player-by-player basis. Obviously the pre-arb years are found money, but don't think arbitration awards are all that good at tracking actual value. How accurately people are paid during those years is going to be wildly divergent, and it won't affect all the players the same way.

For example, Heyward is more like 90% surplus, not 75%, since he was paid for about 2-3 WAR and generated 30. Using 50% of arb years is probably going to consistently make the best players look worse. That's probably fine for certain aggregate purposes, but not if you are intending to compare the players themselves.
 
You must log in or register to reply here.
Back
Top