What constitutes a good draft? My study of 1998-2005

[Frobby]

Just to wrap up a few things, the teams that drafted best were the Cardinals (334.7) and Phillies (309.4).   The worst were the Mariners (70.7), Giants (79.4) and Yankees (99.7).   Just ahead of them, the Orioles at 105.1.

The Phillies should get some special mention here because they had 7 drafts in a row at 25+ rWAR, before finally blowing one in 2005, the last year of my study.   Also, Oakland never had a draft below 15 rWAR during the entire 8 year period.

The Cardinals were the champions largely on the back of their 1998 draft where they chose Albert Pujols in the 13th round.   That was the best single draft any team had in the entire 8 year period.   But you could eliminate that one entirely and they still would have been well above average just using the other 7 years' total.   

As to the Orioles, they basically only had two good drafts, 1999 and 2003.   Those basically boiled down to three players, Brian Roberts (1999), Erik Bedard (1999) and Nick Markakis (2003).  The 1999 draft, though the best of any Orioles draft in that period, was the one where the O's had a ton of first round picks as compensation for free agents they had lost.   Almost none of those picks panned out, but BRob did and we stole Bedard in the 6th round.

Finally, I have to note that not only did the Mariners come in dead last at 70.7 rWAR, but almost half that total came from Adam Jones while he was an Oriole.

[HowAboutThat]

30 minutes ago, Frobby said:

nd.

Finally, I have to note that not only did the Mariners come in dead last at 70.7 rWAR, but almost half that total came from Adam Jones while he was an Oriole.

Please do not share that with the good people at Lookout Landing, ha ha.

I am very grateful for all the work you did. I know that even if you can just find this information, it is very tedious to go through it and you did all of this work 30 times.
Hats off…

[HowAboutThat]

I echo @Picklesrequest for a rough estimate of what a given draft pick should deliver. However that would probably be way too tedious.

Also, I’m not sure it’s germane.

[Pickles]

1 minute ago, Philip said:

I echo @Picklesrequest for a rough estimate of what a given draft pick should deliver. However that would probably be way too tedious.

Also, I’m not sure it’s germane.

I'd say it is germane.  It certainly is a different question.

If Frobby wants to do it, he can knock himself out.  But I bet someone has done that, and we just need to find the study.

[Frobby]

10 minutes ago, Philip said:

I echo @Picklesrequest for a rough estimate of what a given draft pick should deliver. However that would probably be way too tedious.

Also, I’m not sure it’s germane.

Yeah, this has been studied numerous times.   It's a very steep curve that flattens out pretty quickly.  Here's an old article; there are probably newer ones around but I doubt the picture has changed much.   The Baseball Analysts: Draft Picks and Expected Wins Above Replacement

You can get the information pick by pick by looking at the BB-ref draft database.   Here's the page for the overall no. 1 pick; look at the bottom and it tells you how many made the majors and what their average WAR has been.   However, you have to factor in that the more recent picks either haven't finished their careers, or in some cases, haven't made the majors yet.   You can move from the no. 1 pick to the no. 2 pick etc. very easily.    1st Picks Overall in the MLB June Amateur Draft | Baseball-Reference.com

[GuidoSarducci]

56 minutes ago, Pickles said:

I'd say it is germane.  It certainly is a different question.

If Frobby wants to do it, he can knock himself out.  But I bet someone has done that, and we just need to find the study.

Here's what I came up with from years 1990 to 1999 based on baseball reference data

Pick	Avg rWAR
1	30.72
2	10.25
3	5.39
4	11.92
5	-0.43
6	7.39
7	10.38
8	10.09
9	8.55
10	3.09
11	9.76
12	7.50
13	9.97
14	5.71
15	9.55
16	10.08
17	3.88
18	2.37
19	12.35
20	14.45
21	7.77
22	2.45
23	6.11
24	2.82
25	0.24
26	0.79
27	-0.19
28	4.85
29	1.37
30	-0.22

 

Note: I just took the raw data.  There are some anomalies like players being counted twice (e.g, JD Drew)  I''m not sure how that can be dealt with gracefully, however, unless they were picked the same # each time, you could make an argument they should be included. 

Edited August 13, 2021 by GuidoSarducci

[Frobby]

5 minutes ago, GuidoSarducci said:

Here's what I came up with from years 1990 to 1999 based on baseball reference data

Pick	Avg rWAR
1	30.72
2	10.25
3	5.39
4	11.92
5	-0.43
6	7.39
7	10.38
8	10.09
9	8.55
10	3.09
11	9.76
12	7.50
13	9.97
14	5.71
15	9.55
16	10.08
17	3.88
18	2.37
19	12.35
20	14.45
21	7.77
22	2.45
23	6.11
24	2.82
25	0.24
26	0.79
27	-0.19
28	4.85
29	1.37
30	-0.22

This really goes to show what a crapshoot the draft can be.  10 years of drafting and the no. 5 overall pick was in negative territory.   That’s a fluky result if you looked over a longer period, but still…

For the record, the O’s had no #5 picks in that time window!

[NCRaven]

4 minutes ago, Frobby said:

For the record, the O’s had no #5 picks in that time window!

You know, I was wondering.  Thanks for taking the time to do this.

[NCRaven]

The average Hall of Famer runs 50-70 WAR. Maybe our expections for AR are running a little high.

[Pickles]

4 hours ago, GuidoSarducci said:

Here's what I came up with from years 1990 to 1999 based on baseball reference data

Pick	Avg rWAR
1	30.72
2	10.25
3	5.39
4	11.92
5	-0.43
6	7.39
7	10.38
8	10.09
9	8.55
10	3.09
11	9.76
12	7.50
13	9.97
14	5.71
15	9.55
16	10.08
17	3.88
18	2.37
19	12.35
20	14.45
21	7.77
22	2.45
23	6.11
24	2.82
25	0.24
26	0.79
27	-0.19
28	4.85
29	1.37
30	-0.22

 

Note: I just took the raw data.  There are some anomalies like players being counted twice (e.g, JD Drew)  I''m not sure how that can be dealt with gracefully, however, unless they were picked the same # each time, you could make an argument they should be included. 

Thanks for that.  That's great work. It's the basis of an article you should publish.

And to Frobby for the link above.  

What's amazing about this information provided is that it suggests that outside of the top pick, the first twenty are pretty much a crap shoot.  Or at least that's how it played out in the 90s.

One realizes when talking about the MLB amateur draft that there is far more that we can never know than that we can.  That's inherently distasteful to the human mind I believe, particularly the type drawn to the analytical world of baseball.

[NelsonCruuuuuz]

7 hours ago, Frobby said:

Over in the Draft forum, 7Mo asked a few weeks ago what was the median value of a team draft.   I took a guess that it was in the 20-25 rWAR range, but the topic has been eating at me, so I decided to take a close look.

For my study, I used 1998 - 2005.  I used 1998 because that was the first year that the major leagues had 30 teams, and I stopped at 2005 because I figured as I got further beyond that, more and more players would still be active and accruing more WAR.   And so, I looked at every draft of every team for those 8 years, a total of 240 team drafts.   In summary, here is what I found.

1.  The mean average team draft produced 23.2 rWAR.   However, the median team draft produced only 14.5 rWAR.   That's because the good drafts can go way beyond 30 WAR but the bad ones can't really get much below 0 WAR.   

2.  The mean average total value over the 8 year period was 185.3 rWAR, whereas the median was 183.3.   Those two measures line up over a larger sample size.

3.  The frequency of various WAR ranges for a draft looks like this:

< 0: 10.8% (26/240)

0 - 4.9: 16.7 % (40/240)

5 - 9.9: 11.7% (28/240)

10 - 14.9: 11.3% (27/240)

15 - 19.9: 5.8% (14/240)

20 - 24.9: 6.7% (16/240)

25 - 29.9: 4.6% (11/240)

30 - 39.9: 8.8% (21/240)

40 - 49.9 10.0% (24/240)

50+: 13.8% (33/240)

So, bottom line, if your team wants to produce an above average pipeline, you'd better not have too many drafts that produce below 14.5 rWAR, you'll need to average better than 23.2 rWAR per draft, and you'll need to spring a 50+ rWAR draft once every 7-8 years or better.   

Here are a few qualifiers:

1.   This was a study of 1998-2005.   The draft rules have changed a number of times since then and that may have affected how rWAR is distributed.   However, I tend to doubt it has had significant affect.   

2.  Needless to say, draft order affects how well a team drafts, on average.

3.  The numbers I used were career WAR for players a team drafted and signed.   I did not factor in how much of that WAR was earned while the player was with the drafting team (which would have been incredibly time consuming to do), or while the player theoretically was under team control (ditto).   I didn't give teams credit for players they drafted but didn't sign.   I didn't factor in whether a team had lost or gained picks under the various rules about FA compensation, etc.   I can say that there were many good players drafted and signed who earned most or all of their WAR with teams other than the one that signed them.   

4.  Not all drafts are equal.   The best drafts I looked at (1998 and 2001) produced about 50% more total WAR than the worst ones (2000 and 2002).   So obviously, what constitutes a good draft for a team can vary according to the talent level available in that draft.

Anyway, I hope folks will find this a useful yardstick when we evaluate Orioles drafts past and future.   Of course, it takes a long time for all the results of a draft to come in.   

In my next post I will post a spreadsheet of all the team results over the 8 year period.   It goes without saying that it's not pretty for the Orioles during that time period.

This is amazing Frobby!!!! Thank you!!!

[Pickles]

4 hours ago, NCRaven said:

The average Hall of Famer runs 50-70 WAR. Maybe our expections for AR are running a little high.

Basically, it seems to me, half the board is going to be disappointed if he isn't a HOFer.  That's setting yourself up to be disappointed.  If he is Wieters with better plate discipline, and has a knack for not getting injured in his peak controllable years, I'll be thrilled.  

[OsFanSinceThe80s]

1 minute ago, Pickles said:

Basically, it seems to me, half the board is going to be disappointed if he isn't a HOFer.  That's setting yourself up to be disappointed.  If he is Wieters with better plate discipline, and has a knack for not getting injured in his peak controllable years, I'll be thrilled.  

My hope is AR is more of .280 type hitter that racks up doubles and hits around 20 home runs a season. 

[Pickles]

Just now, OsFanSinceThe80s said:

My hope is AR is more of .280 type hitter that racks up doubles and hits around 20 home runs a season. 

Without changes to the game, I'm not sure he's a .280 hitter.

I'll be thrilled with Wieters with better plated discipline:

250/350/450

With good defense behind the plate, that is a ~4 win player, depending on how hard you ride them.

[Frobby]

Philip’s question got me curious how much of the draft value comes from the early rounds.   It’s a lot, but it varies a ton from year to year, and there’s plenty of value down below too.

Overall, during the 8 year period, 54.6% of the value in the draft came from the first three rounds: 38.5% from the first round, 10.7% in the second round, 5.4% in the third round.   That still leaves 45.4% of the value in the draft coming after the third round, so plenty of nuggets there.   In the two most “orderly” years, 2003 and 2005, 71.5% and 70.3% of the value came in the first three rounds, and both of those were very strong drafts.   In the least “orderly” draft, 2000, only 34.4% of the value came from the first three rounds.   That was the only year where the value in the first three rounds was under 50% of total value, so it’s a huge outlier.  

The best first round during this period was 2005, which produced a whopping 406.4 rWAR, about 54.4% of the total value of that year’s draft.   The average 1st round pick (there were 48 of them) was worth 8.5 rWAR.   Of course the Orioles, picking unlucky 13th, chose Brandon Snyder, worth -0.1 rWAR, and then closed the supplemental portion of the first round by choosing Garret Olson, worth -2.7 rWAR.    In a year when making first round picks was like shooting fish in a barrel, the O’s couldn’t even find the barrel.

The weakest first round in this period was 2000, worth only 174.9 rWAR, 25.0% of the value in that draft.  The O’s did their part by using the 14th overall pick on Beau Hale, who never made the majors.  He was far from alone, as 14 of the top 28 never made the Show, including six players drafted ahead of Hale.   Incredibly, the second round that year was even worse (relatively speaking) than the first, producing only 15.3 rWAR, vastly lower than any other year I looked at.  Only four second rounders that year topped 1.0 rWAR, led by Chad Qualls at 5.8.   The O’s had no second rounder that year, having forfeited it to sign the studly Mike Trombley.  They can take solace in knowing that the player the Twins selected in their spot, J.D. Durbin, turned out to be worth -0.5 rWAR.   The interesting thing about the weak 1st/2nd round is that 2000 wasn’t a bad draft overall; it was just about average at 699.8 rWAR.  A lot of talent that teams missed in the early rounds that year.  I count nine players who produced at least 20 rWAR in their career, including Cliff Lee (43.2),Yadier Molina (41.3), Jose Bautista (36.7), Edwin Encarnacion (35.4), Brandon Webb (31.1), James Shields (31.1), Mike Napoli (26.3), Jason Bay (24.8) and David DeJesus (23.4).   Tons of guys in the 10-20 rWAR range too.

The 2002 draft, in a stacked year, produced more value in the second round (175.9 rWAR) than the 2000 draft did in the first (174.9), even though 353.8 rWAR worth of talent already had been snapped up in the 2002 first round.  That second round featured Joey Votto (63.0 rWAR), Jon Lester (43.3) and Brian McCann (31.9), along with 14 others who signed and reached the majors.  The Orioles’ no. 2 pick, Corey Shafer, wasn’t one of them.

My overall takeaway is that the first three rounds are very important but the talent varies a lot from year to year, and some years are easier than others to identify where the talent is.