35 replies to this topic

[Eric Van]

The Sox have promoted guys so aggressively that it's hard to get a handle on how guys are doing. Clay Davenport's Peak Translations are the most ambitious effort I know of to adjust for age relative to league, so I always find them worth looking at.

He has very clearly changed his Peak Translation formulas since last year; while they used to be too rosy for guys in general, they are now progressively more pessimistic as you go lower in the minors. Will Middlebrooks projects to have a peak TAv (Eqa) of .257, which is below MLB average -- but he's 4th in the league after Eric Hosmer, Lavarnway, and Derek Norris.

So I've re-scaled all the AA, A+, and A numbers to match the AAA numbers (where the average player projects to have a position-adjusted .232). And I've added a positional adjustment. Here's the very interesting ranking, which doesn't include defense or luck (two tweaks I may make later).

The numbers appear to be a few days out of date, but they give you an idea of what to make of some of the tough-to-read guys.

Kalish is ranked as a RF; as a CF he's .309. (Hazelbaker as a CF would be .274, and he can definitely play there.)

Davenport Peaks, Position-Adjusted

Name	Peak
Ryan_Kalish	.298
Ryan_Lavarnway	.298
Will_Middlebrooks	.282
Lars_Anderson	.281
Anthony_Rizzo	.276
Yamaico_Navarro	.271
Oscar_Tejeda	.268
Luis_Exposito	.265
Jeremy_Hazelbaker	.263
Michael_Almanzar	.261
Reymond_Fuentes	.261
Daniel_Butler	.260
Alex_Hassan	.257
Jose_Iglesias	.256
Daniel_Nava	.256
Che-Hsuan_Lin	.255
Christopher_McGuiness	.252
Nate_Spears	.250
Tim_Federowicz	.249
Peter_Hissey	.248
Derrik_Gibson	.246
Mitch_Dening	.242
Christian_Vazquez	.242
Ronald_Bermudez	.238
Niuman_Romero	.237
Ray_Chang	.231
Chih-Hsien_Chiang	.230
Ryan_Dent	.229
Bubba_Bell	.225
Dusty_Brown	.222
Josh_Reddick	.221
Jorge_Padron	.217
Mark_Wagner	.215
Zach_Gentile	.213
Gil_Velazquez	.212
Tug_Hulett	.210
Jonathan_Hee	.203
Reynaldo_Rodriguez	.203
Aaron_Bates	.200
Shannon_Wilkerson	.197
Jorge_Jimenez	.188
Drew_Hedman	.188
Vladimir_Frias	.117

Some Lowell numbers:

Davenport Peaks, Position-Adjusted

Name	Peak
Jose_Garcia	.282
Miles_Head	.262
Kolbrin_Vitek	.260
Felix_Sanchez	.260
Brandon_Jacobs	.258
Bryce_Brentz	.202
Joantoni_Garcia	.186
David_Renfroe	.179

Edited by Eric Van, 07 September 2010 - 04:43 PM.

[ELMER FLICK]

Fascinating stuff. Do you have the Davenport rankings before adjustment? It would be interesting to compare.

Looking through the rankings, most of them are not too far from what I would have expected. Three surprised me. I knew Hazelbaker was having a good season, but I had no idea it was that good. Iglesias, on the other hand, seemed to be doing well at the plate, better than most scouts expected, yet even with a position adjustment, he is only middle of the pack. And yet he is one place ahead of Nava, the other surprise.

The more aggressive promotion strategy also intrigues me. Is it a lesson learned from Clay Buchholz? I've seen several times in the press lately the assertion that, because he had never struggled in the minor leagues, he had a lot more trouble adjusting when he faced adversity in MLB. Or is it because they have so many high upside prospects that they feel compelled to challenge them, to get a quicker read on which are likely to succeed?

[Eric Van]

Fascinating stuff. Do you have the Davenport rankings before adjustment? It would be interesting to compare.

They're available at BP, here.

Here's an interesting alternative ranking (with most guys below .240 omitted), which takes 1/3 (arbitrary guess) of the luck on balls in play as estimated by MinorLeagueSplits.com and adds or subtracts it to the above rankings. (It uses the estimation formula TAv = .4605 * OBP + .2254 * SA + .0151.)

Navarro is ranked as a 3B here, rather than as a SS (he loses .015).

Davenport Peaks, Position and Luck-Adjusted

Name	Peak	OBP Lu	SA Lu	Adj
Ryan_Kalish	.298	.007	.009	.300
Ryan_Lavarnway	.298	-.004	-.005	.297
Anthony_Rizzo	.276	.031	.048	.285
Lars_Anderson	.281	-.014	-.023	.277
Will_Middlebrooks	.282	-.035	-.048	.273
Luis_Exposito	.265	.018	.026	.270
Jose_Garcia	.282	-.048	-.069	.270
Miles_Head	.262	.017	.019	.266
Michael_Almanzar	.261	.014	.018	.264
Daniel_Nava	.256	.033	.048	.264
Che-Hsuan_Lin	.255	.017	.024	.260
Yamaico_Navarro	.256	.010	.014	.259
Nate_Spears	.250	.028	.044	.258
Oscar_Tejeda	.268	-.044	-.060	.257
Alex_Hassan	.257	-.003	-.003	.257
Jeremy_Hazelbaker	.263	-.025	-.042	.256
Daniel_Butler	.260	-.016	-.023	.256
Reymond_Fuentes	.261	-.023	-.029	.255
Christopher_McGuiness	.252	-.005	-.006	.251
Brandon_Jacobs	.258	-.031	-.044	.250
Kolbrin_Vitek	.260	-.040	-.066	.249
Tim_Federowicz	.249	-.003	-.004	.249
Jose_Iglesias	.256	-.045	-.054	.245
Derrik_Gibson	.246	-.007	-.011	.244
Peter_Hissey	.248	-.028	-.037	.241
Felix_Sanchez	.260	-.081	-.102	.240
Bryce_Brentz	.202	.055	.096	.217

[SoxScout]

Awesome stuff. What does Lavarnway look like as a 1B/DH?

[Eric Van]

Awesome stuff. What does Lavarnway look like as a 1B/DH?

Catchers moving to 1B lose .034. He's just another guy as a 1B, an elite prospect as a C.

Here's the adjustment table, which is based on the TrA of regular players, not everyone who plays the position:

C +.017
1B -.017
2B +.004
3B -.001
SS +.014
LF -.008
CF +.006
RF -.005
DH -.017

[Eric Van]

As of sometime in the last week or two, Clay completely revised or corrected his algorithms, as the peak projections of guys who had already been promoted from a level changed by more than the usual .001. It's possible that the projections have been merely improved via some new insight, but (since the algorithm changed a ton from last year) it's also possible that there was a bug in the earlier version.

Despite my curiosity I am resisting the temptation to completely redo my spreadsheet since the season will be over in a week. But look for final numbers on September 7th, probably early in the morning.

In the meantime, though, Rizzo is leading the EL (non-position adjusted) after being 2nd in the CL, where Tejeda is now 2nd. Fuentes is 7th in the SAL. Kalish and Lavanway are top 10 guys, too.

[Eric Van]

Here are the final numbers, and I find them utterly fascinating (hence the change to the thread title).

The question of properly adjusting for age relative for ml players to level is one of the thorniest in sabermetrics. How much air should we take out of Jeremy Hazelbaker's numbers? How much pizazz should we give to Reymond Fuentes instead?

Clay Davenport's Peak Translations are the only comprehensive attempt that I know of to answer those questions. Of course, they should not be taken as gospel, but they should absolutely be used to inform our prospect ratings.

His raw numbers are available at Baseball Prospectus (under Minor League True Average) and you can find stats for the NYP and GCL, even though they are not linked to on the page, by hacking the URL (you can also put in a team code such as BOS to get all the stats for that team).

I have made two adjustments to Clay's numbers.

First, the average peak projection in Clay's system drops as you go down the minors. It's .236 in the IL, .217 in the EL, .198 in the CRL, .186 in the SAL, .171 in the NYPL, and .161 in the GCL. That reflects the reality that guys farther from the bigs are less juicy as prospects than guys in the high minors, but we include that knowledge when we rank prospects anyway. We would really like a set of numbers where the best player in any league had the same peak projection as the best in any other league regardless of level, and ditto for the average player. That tells you how the guy hit relative to his peer group, age and level adjusted. So I have normalized every league to the .236 TAv of the IL.

Second, I've added a positional adjustment: +.017 for C, +.014 for SS, +.006 for CF, +.004 for 2B, -.001 for 3B, -.005 for RF, -.008 for LF, and -.017 for 1B. These are based on the average TAv of MLB regulars at each position.

Here, without comment, are the rankings. To say there are some surprises is an understatement (Pete Hissey lives!)

Final Position and Level-Adjusted Davenport Peak Hitting Translations

Name	Age	Tm	Pn	PA	TAv
Ryan_Kalish	22	BOS	CF	447	.302
Ryan_Lavarnway	22	PME	C	550	.297
Reymond_Fuentes	19	GRN	CF	420	.295
Anthony_Rizzo	20	PME	1B	626	.293
Jose_Vinicio	16	GCL	SS	172	.291
Oscar_Tejeda	20	SLM	2B	560	.290
Jose_Garcia	19	GRN	SS	220	.283
Keury_De La Cruz	18	GCL	CF	224	.282
Michael_Almanzar	19	GRN	3B	542	.282
Lars_Anderson	22	BOS	1B	543	.279
Yamaico_Navarro	22	BOS	SS	459	.277
Che-Hsuan_Lin	21	PME	CF	555	.276
Daniel_Butler	23	SLM	C	382	.276
Will_Middlebrooks	21	SLM	3B	485	.274
Jose_Iglesias	20	PME	SS	284	.273
Peter_Hissey	20	SLM	CF	511	.271
Christian_Vazquez	19	GRN	C	303	.268
Luis_Exposito	23	PME	C	536	.267
Felix_Sanchez	20	GRN	CF	254	.266
Josh_Reddick	23	BOS	CF	508	.266
Brandon_Jacobs	19	LOW	LF	267	.261
Kolbrin_Vitek	21	GRN	3B	287	.259
Jeremy_Hazelbaker	22	GRN	CF	511	.259
Alex_Hassan	22	SLM	LF	409	.257
Nate_Spears	25	PME	2B	605	.256
Daniel_Nava	27	BOS	RF	456	.253
Derrik_Gibson	20	GRN	SS	567	.253
Jason_Thompson	19	GCL	2B	193	.253
Henry_Ramos	18	LOW	RF	184	.252
Chih-Hsien_Chiang	22	PME	RF	480	.252
Ryan_Dent	21	SLM	SS	427	.251
Sean_Killeen	23	GRN	C	123	.250
Heiker_Meneses	18	GCL	3B	227	.250
Josue_Peley	22	LOW	C	116	.247
Mitch_Dening	21	SLM	RF	508	.247
Tim_Federowicz	22	SLM	C	456	.246
Ray_Chang	26	PME	3B	497	.244
Ronald_Bermudez	22	SLM	CF	480	.242
Miles_Head	19	LOW	1B	273	.242
Bubba_Bell	27	PAW	RF	392	.240
Dusty_Brown	28	PAW	C	288	.233
Wilfred_Pichardo	20	GRN	RF	242	.233
Gustavo_Molina	28	PAW	C	126	.232
David_Renfroe	19	LOW	3B	227	.230
Luke_Yoder	22	GCL	LF	196	.229
Moko_Moanaroa	20	GCL	LF	122	.228
Joantoni_Garcia	19	LOW	2B	236	.227
Kenneth_Roque	20	GRN	2B	188	.226
Oscar_Perez	18	GCL	C	91	.225
Seth_Schwindenhamm	18	LOW	LF	123	.223
Mark_Wagner	26	PAW	C	151	.221
Reynaldo_Rodriguez	24	GRN	1B	328	.220
Gil_Velazquez	30	PAW	SS	259	.218
Jorge_Padron	23	PME	LF	327	.216
Aaron_Bates	26	PAW	LF	489	.216
James_Kang	22	LOW	2B	165	.215
Jason_Place	22	SLM	RF	110	.214
Shannon_Wilkerson	21	GRN	LF	294	.214
Bryce_Brentz	21	LOW	RF	288	.214
Ryan_Khoury	26	PME	2B	317	.213
Niuman_Romero	25	PAW	2B	388	.212
Jonathan_Hee	24	SLM	2B	315	.207
William_Holmes	22	GCL	RF	144	.206
Luis_Segovia	23	SLM	SS	136	.206
Zach_Gentile	23	GRN	2B	381	.205
Matt_Sheely	23	PME	RF	270	.205
Trygg_Larsson-Danforth	23	GCL	1B	170	.203
David_Mailman	21	SLM	LF	93	.193
Jorge_Jimenez	25	PAW	3B	370	.186
Vladimir_Frias	23	GRN	SS	333	.184
Drew_Hedman	23	SLM	1B	353	.181
William_Vazquez	25	PME	C	163	.180
Chia-Chu_Chen	21	LOW	C	86	.164

Itty Bitty Sample Size Appendix

[td=*][color="#000000"]C[/color] [/td][td=*][color="#000000"]45[/color] [/td][td=*][color="#000000"].224[/color] [/td][td=*][color="#000000"]Jayson_Hernandez [/color] [/td][td=*][color="#000000"]21[/color] [/td][td=*][color="#000000"]LOW [/color] [/td][td=*][color="#000000"]C[/color] [/td][td=*][color="#000000"]68[/color] [/td][td=*][color="#000000"].215[/color] [/td][td=*][color="#000000"]Juan_Carlos_Linares[/color] [/td][td=*][color="#000000"]25[/color] [/td][td=*][color="#000000"]PME [/color] [/td][td=*][color="#000000"]RF[/color] [/td][td=*][color="#000000"]47[/color] [/td][td=*][color="#000000"].195[/color] [/td][td=*][color="#000000"]Nick_Robinson [/color] [/td][td=*][color="#000000"]22[/color] [/td][td=*][color="#000000"]LOW [/color] [/td][td=*][color="#000000"]2B[/color] [/td][td=*][color="#000000"]70[/color] [/td][td=*][color="#000000"].193[/color] [/td][td=*][color="#000000"]Bryan_Peterson [/color] [/td][td=*][color="#000000"]20[/color] [/td][td=*][color="#000000"]PME [/color] [/td][td=*][color="#000000"]RF[/color] [/td][td=*][color="#000000"]61[/color] [/td][td=*][color="#000000"].190[/color] [/td][td=*][color="#000000"]Roberto_Ramos [/color] [/td][td=*][color="#000000"]21[/color] [/td][td=*][color="#000000"]GCL[/color] [/td][td=*][color="#000000"]RF[/color] [/td][td=*][color="#000000"]42[/color] [/td][td=*][color="#000000"].155[/color] [/td][td=*][color="#000000"]Trey_Lucas [/color] [/td][td=*][color="#000000"]22[/color] [/td][td=*][color="#000000"]GCL[/color] [/td][td=*][color="#000000"]C[/color] [/td][td=*][color="#000000"]3[/color] [/td][td=*][color="#000000"].149[/color] [/td][td=*][color="#000000"]Zach_Kapstein [/color] [/td][td=*][color="#000000"]18[/color] [/td][td=*][color="#000000"]GCL[/color] [/td][td=*][color="#000000"]C[/color] [/td][td=*][color="#000000"]9[/color] [/td][td=*][color="#000000"].149[/color] [/td][td=*][color="#000000"]Drew_Dominguez [/color] [/td][td=*][color="#000000"]23[/color] [/td][td=*][color="#000000"]SLM [/color] [/td][td=*][color="#000000"]2B[/color] [/td][td=*][color="#000000"]41[/color] [/td][td=*][color="#000000"].081[/color] [/td]

Name	Age	Tm	Pn	PA	TAv
Leonel_Escobar	19	SLM	C	77	.276
Kendrick_Perkins	18	RSX	RF	9	.271
Sean_Coyle	18	GCL	2B	12	.271
Maykol_Sanchez	22	GCL	C	28	.257
Adalberto_Ibarra	23	SLM	C	53	.256
Boss_Moanaroa	18	GCL	1B	69	.250
Jordan_Sallis	21	GCL	2B	20	.230
Roberto_Reyes	19	GCL	3B	23	.226
Carson_Blair	20	LOW

OK, I lied: I can't help noting that Henry Ramos was .279 before his promotion.

Next I'm going to see if minorleaguesplits.com's luck adjustments actually do what they're designed to do, and if they do, I'm going to toss those into the mix.

Edited by Eric Van, 07 September 2010 - 05:03 PM.

[Detts]

Eric,

Did you consider Vitek as a Greenville guy or a Lowell guy? Only 40 of his AB's were in Greenville this year.

[Eric Van]

Eric,

Did you consider Vitek as a Greenville guy or a Lowell guy? Only 40 of his AB's were in Greenville this year.

Oops, meant to explain that: all numbers combine all play at all levels with two exceptions: rehab assignments are only included if they would improve the overall numbers, and brief super-promotions to cover injury gaps like Dan Butler's are excluded. The listed team is simply the team the player finished the season with.

Vitek's .259 is .258 with Lowell and .266 with Greenville. Kalish was .310 in AA, .297 in AAA, .294 with Boston. Lavarnway was .284 with Salem and .320 with Portland. Rizzo was .280, then .297, while Lars was .304, then .276. I already mentioned Henry Ramos, whose .087 in 26 PA with Lowell dragged his number from .279 down to .252.

[someoneanywhere]

For me, there's no one bottom-feeding who surprises me -- but there are a couple on the top end. If Dan Butler is TAVing at that tune, he's a prospect. He is already an advanced defensive catcher -- ahead of both Expo and Lavarnway and Fedex in my book -- and if you buy the conventional wisdom that catching prospects mature later than other every day prospects, I think you have to consider him in your mix. The other is Almanzar. As most of you know I'm pro-Mike. But most of that was projection and tools and, especially, the chance for big-time power. That there is some basis for hope beyond what the eyes see is encouraging. The same with Fuentes. You can see it, and you can salivate on it, but it's something else to note that the translations have him pegged that high.

[Koufax]

Incredibly interesting that Jose Iglesias is not ranked higher. Not having paid attention to Reymond Fuentes, I will start doing so now. Thanks for presenting this data.

[Eric Van]

Incredibly interesting that Jose Iglesias is not ranked higher. Not having paid attention to Reymond Fuentes, I will start doing so now. Thanks for presenting this data.

The thing that makes these translations interesting is that Clay has looked at how the separate hitting components project. BABIP doesn''t project well. Iso and IsoD, especially at young ages, project well, and I suspect speed does as well. Iglesias is merely good here because his numbers are very BABIP-dependent.

In their first years these translations (IIRC) had Lavarnway way, way ahead of Fed-X even though the latter had better numbers, playing at the same age for the same team. Ryan had the Iso and IsoD, Fed-X was all BA. It took exactly one year for those translations to come up bingo.

I wonder whether this latest version overrates youth -- Fuentes, Hissey, and Almanzar are all getting big boosts for being young for their level. But as I said they are the only attempt to objectively determine how much of a boost should be given for that, and I think it's safe to say that even if the translations are over-adjusting, that most prospect raters have at least Almanzar and Hissey too low. I have them at 31-32 and 38-40 respectively in the rankings I'm working on whereas SP has them at 43 and 53.

[Eric Van]

For me, there's no one bottom-feeding who surprises me -- but there are a couple on the top end. If Dan Butler is TAVing at that tune, he's a prospect. He is already an advanced defensive catcher -- ahead of both Expo and Lavarnway and Fedex in my book -- and if you buy the conventional wisdom that catching prospects mature later than other every day prospects, I think you have to consider him in your mix.

SP has just bumped him up to 39, and if that's not right it's only a few notches too high.

"This guy's been unbelievable – the way he's run the pitching staff, his energy, the confidence he has, his maturity even for his inexperience in professional baseball," said Hazen. "He's a good sign. There are a lot of guys who walk out of there and say, 'I don't know when or how, but Dan Butler is going to play in the big leagues someday.'"

"Look at Daniel Nava. He went undrafted. You look at Dan Butler, a kid who played at Arizona but never played, backed up, and now he's getting an opportunity to play and look what he's doing. He's played himself into being a prospect. These stories motivate the hell out of people and it's awesome to see," said Red Sox minor league catching instructor Chad Epperson. "At some point in time, [Butler] will be a big leaguer."

Full story,

[Detts]

Butler and Nava are a good comparison.

If I remember correctly, Butler is listed at 5'10", however I'm guessing that height was measured in his cleats on a sidewalk, and then rounded up.

He's another guy trying to overcome the 'you have to be big to be a catching prospect' hurdle.

[Eric Van]

Next I'm going to see if minorleaguesplits.com's luck adjustments actually do what they're designed to do, and if they do, I'm going to toss those into the mix.

Well, I've only got 50 pairs of player seasons in the database so far, but it's very clear that "luck" (the difference between BABIP and expected BABIP based on GB/ LD / FB breakdown) for minor leaguers is largely a skill.

The correlation from year to year for OBP is .43.

The correlation for "luckiness" on OBP (luck-adjusted OBP minus real OBP -- it's going to be the same as the BA differential) is ... .44. Even with just 50 pairs of seasons that has a 1 in 1000 chance of being random.

However, as you might expect, there does appear to be some luck that's being measured. If you want to predict a player's OBP it seems like you can do a better job by tossing in some of last year's luck differential. I need to load up quite a bit more data before I can even make a guess as to how much of the luck to count, though.

The really interesting question, though, is this: what does it mean if a player consistently has a higher or lower BABIP than expected?

Tejeda has been consistently better than expected ("lucky"), ditto Middlebrooks. Ellsbury was hugely "lucky."

Lin is the guy most consistently underperforming, also Almanzar. I looked at a handful of guys who have disappointed in MLB and most of them were "lucky" as prospects.

My guess is that being consistently "lucky" is a good sign, which is what you'd expect, since it means that your line drives, fly balls, and grounders are as a whole harder than ml average.. The question is whether being neutral or "unlucky" (Kalish, Rizzo and Lavarnway the last two years) is any kind of warning sign.

[Detts]

Well, I've only got 50 pairs of player seasons in the database so far, but it's very clear that "luck" (the difference between BABIP and expected BABIP based on GB/ LD / FB breakdown) for minor leaguers is largely a skill.

The correlation from year to year for OBP is .43.

The correlation for "luckiness" on OBP (luck-adjusted OBP minus real OBP -- it's going to be the same as the BA differential) is ... .44. Even with just 50 pairs of seasons that has a 1 in 1000 chance of being random.

However, as you might expect, there does appear to be some luck that's being measured. If you want to predict a player's OBP it seems like you can do a better job by tossing in some of last year's luck differential. I need to load up quite a bit more data before I can even make a guess as to how much of the luck to count, though.

The really interesting question, though, is this: what does it mean if a player consistently has a higher or lower BABIP than expected?

Tejeda has been consistently better than expected ("lucky"), ditto Middlebrooks. Ellsbury was hugely "lucky."

Lin is the guy most consistently underperforming, also Almanzar. I looked at a handful of guys who have disappointed in MLB and most of them were "lucky" as prospects.

My guess is that being consistently "lucky" is a good sign, which is what you'd expect, since it means that your line drives, fly balls, and grounders are as a whole harder than ml average.. The question is whether being neutral or "unlucky" (Kalish, Rizzo and Lavarnway the last two years) is any kind of warning sign.

Could it be speed? In the minors defense overall isn't as good as the majors, so you would expect to turn more balls into hits if you can run fast...

[Eric Van]

Could it be speed? In the minors defense overall isn't as good as the majors, so you would expect to turn more balls into hits if you can run fast...

Lin is fast, so I don't think it's speed.

Edit: this isn't quite right but I have to go eat a yummy Mom-cooked New Year's meal ... more later.

Looking at the player seasons so far ... hitters with high OBP have a very big tendency to have been "unlucky." So hitting the ball harder is simply something better hitters do.

So the cool question is whether we can use this to identify the actual lucky or unlucky guys. A good hitter with a lower than expected BABIP may have actually been unlucky.

The crappy thing is that I didn't put BA into this database and now I wish I had so that I could look at it and Iso separately. I may have to look up my 74 seasons again ....

And of course if I wanted to put in the separate LD, GB, FB data that would be even more informative, but that's a ton of work.

Edited by Eric Van, 08 September 2010 - 03:57 PM.

[Buzzkill Pauley]

Could it be speed? In the minors defense overall isn't as good as the majors, so you would expect to turn more balls into hits if you can run fast...

I think the most valuable thing would be to see where the "luck" occurs vis-a-vis FB/GB/LD falling for hits.

If the luck is in line drives, then I can see the luck being a "hits the ball hard" type of thing. I don't believe one could easily make the case for BABIP data to reflect a "hit it where they ain't" level of superior bat control, as each batter is working on specific skills in the minor leagues, as well as trying to put up good numbers. So I would tend to think that luck on grounders and fly balls above the norm is reflective of, well, luck. Perhaps speed if it's a boost on grounders.

I tend to thin this is like fielding stats, where a greater sample size than a single season is needed to derive a reasonable projection from the data. To me that makes sense because there's a relatively limited number of contact events once they're broken down by type.

Or, frankly, it could be reflecting something totally different, like how much additional topspin the player's swing produces upon contact, causing line drives to sink and grounders to take weird bounces. If that's the case, then it definitely reflects a skill.

[Eric Van]

Well, looking at the data more closely just leads to more confusion and a realization that I need much more data to make sense of it all.

The guys with the highest "expected SA" according to batted ball type (but which still counts all HR as HR, mind you) tend to have lower actual OBP and SA, that is, to underperform BA on balls in play. Guys with very low "expected SA" tend to have higher actual OBP and SA.

Looking at the 16 seasons in my database (so far) with the highest expected SA, 12 of them were "unlucky." Looking at batted ball type BABIP, the general tendency for these 16 guys is to be "lucky" on ground balls, "lucky" on line drives, and -- the one near constant -- very "unlucky" on fly balls. I would say that guys who hit the ball hard hit lots of home runs, tend to hit harder grounders and line drives, but hit a lot fewer bloop fly ball hits than guys with less power. Now, the system in use here does not report the same expected BABIP for everyone each year so it is already attempting to factor in some other elements. But clearly it could do a much better job.

The other thing that seems to be true is that a player's "luck" profile seems to be part of his makeup. Josh Reddick has been "lucky" on GB and "unlucky" on FB four straight years and "unlucky" on line drives three years out of four.

Here are the 16 seasons, the overall unlucky ones in red. Unfortunately, I coded the luck by batted ball type just + or - no matter how small the margin was (I should have coded 0 for small margins; Exposito's positive luck on FB is just 2 points).

Luck of Top Expected Sluggers

Name	Year	Exp	Act	Luck
Reddick	07	541	526	G+F-L-
Lavarnway	09	540	537	G+F-L-
Reddick	08	538	546	G+F-L-
Kalish	10	527	502	G-F-L+
Rizzo	10	527	477	G-F-L+
Lavarnway	10	512	503	G+F-L+
Anderson	08	507	520	G+F+L+
Exposito	08	507	508	G-F+L+
Reddick	10	502	461	G+F-L-
Nava	10	490	459	G+F-L-
Nava	09	488	533	G+F+L+
Chiang	09	486	418	G-F-L+
Reddick	09	485	446	G+F-L+
Spears	10	485	440	G+F-L-
Chiang	10	484	423	G-F-L+
Lin	07	477	411	G-F-L+

I'll look at the "lucky" non-sluggers next.

[Eric Van]

I've now got a spreadsheet with every player in the minors. And I've discovered that, while my instincts were correct in thinking that the published Davenport Peak Transalstions were too tough on players in the lower minors, my method of correcting for this (assuming that the prospects at every level are equally good) actually over-corrected. If you take the published translations verbatim, players that were promoted a level tended to gain, on average, .006 of Peak Projection, which is clearly not right. With my fix, though, they tended to lose -.009. Taking 60% of Clay's figure and 40% of my figure yields a Peak Projected TAv which doesn't change when you get promoted, which is as it should be.

I haven't added my positional adjustments yet, so here are the top 20 hitting prospects in the minors regardless of position (I've thrown in the positional adjustment just FYI):

Top 20 Hitting Prospects, Corrected Davenport Peak Translations

Name	TAv	Org	Pos	Adj	Lvl
Jesus_Montero	.336	NYA	DH?	.319	AAA
Freddie_Freeman	.325	ATL	1B	.308	AAA
Eric_Hosmer	.322	KC	1B	.305	A+/AA
Mike_Moustakas	.307	KC	3B	.306	AAA/AA
Domonic_Brown	.304	PHI	RF	.302	AA/AAA/MLB
Nicholas_Weglarz	.304	CLE	LF	.296	AAA/AA
Anthony_Rizzo	.302	BOS	1B	.285	AA/A+
Lars_Anderson	.295	BOS	1B	.278	AAA/AA
Ryan_Kalish	.293	BOS	CF	.299	AA/AAA/MLB
Alex_Liddi	.292	SEA	3B	.291	AA
Brandon_Belt	.291	SF	1B	.274	A+/AA/AAA
Chris_Marrero	.290	WAS	1B	.273	AA
Greg_Halman	.288	SEA	CF	.294	AAA
Chris_Carter	.287	OAK	1B	.270	AAA
William_Myers	.287	KC	C	.304	A/A+
Mike_Trout	.287	LAA	CF	.293	A/A+
Matt_Dominguez	.286	FLA	3B	.285	AA
Wilmer_Flores	.286	NYN	SS	.300	A/A+
Brett_Jackson	.286	CHN	CF	.292	A+/AA
Lonnie_Chisenhall	.285	CLE	3B	.284	AA

[jon abbey]

Montero caught 105 games this year and DHd 18, he's a catcher right now no matter where the consensus has him ending up and he should get the positional adjustment to match.

[Eric Van]

Montero caught 105 games this year and DHd 18, he's a catcher right now no matter where the consensus has him ending up and he should get the positional adjustment to match.

He still had the most impressive season in the minors either way, though. I'm close to having a full list for that.

[jon abbey]

I know, I just think he should get full credit for how impressive it was.

[Eric Van]

The ranked list of Sox prospects I posted earlier is not wrong. But it doesn't mean what I thought it meant.

It was a list of how well each player had performed, given his age and position, relative to other players at his level. So Fuentes indeed had a terrific year for a CF that age in low-A.

What it did not factor into consideration is that prospects are winnowed out as you move up the minor leagues, so that being the 5th best hitting prospect in your low-A league is not as impressive as being the 5th-best one in your AA league.

This (still very interesting!) list factors that in. It is just Clay's Peak Translations, corrected for the demonstrable bias against higher levels so that guys no longer become better prospects as soon as they're promoted (if anything, they should drop a tiny bit since the promoted guys were a bit likelier to have been lucky), plus the position adjustment.

The first column shows the rank among all prospects in the minors, for those who made the top 100 (a ranking that excluded AAA vets 26 and older, 27 for C, and required 200 AB+BB minimum).

Sox Position-Adjusted (Corrected) Davenport Peak Translations

[td=*].167 [/td][td=*]2B [/td][td=*].171 [/td][td=*] [/td][td=*]Oscar_Perez [/td][td=*]18 [/td][td=*]RSX [/td][td=*]91 [/td][td=*].153 [/td][td=*]C [/td][td=*].170 [/td][td=*] [/td][td=*]David_Mailman [/td][td=*]21 [/td][td=*]SLM [/td][td=*]93 [/td][td=*].172 [/td][td=*]LF [/td][td=*].164 [/td][td=*] [/td][td=*]Josue_Peley [/td][td=*]22 [/td][td=*]PME [/td][td=*]209 [/td][td=*].145 [/td][td=*]C [/td][td=*].162 [/td][td=*] [/td][td=*]William_Vazquez [/td][td=*]25 [/td][td=*]PME [/td][td=*]163 [/td][td=*].142 [/td][td=*]C [/td][td=*].159 [/td][td=*] [/td][td=*]Jonathan_Hee [/td][td=*]24 [/td][td=*]SLM [/td][td=*]315 [/td][td=*].175 [/td][td=*]1B [/td][td=*].158 [/td][td=*] [/td][td=*]William_Holmes [/td][td=*]22 [/td][td=*]RSX [/td][td=*]144 [/td][td=*].158 [/td][td=*]RF [/td][td=*].153 [/td][td=*] [/td][td=*]Trygg_Larsson-Danf [/td][td=*]23 [/td][td=*]RSX [/td][td=*]170 [/td][td=*].170 [/td][td=*]1B [/td][td=*].153 [/td][td=*] [/td][td=*]Drew_Hedman [/td][td=*]23 [/td][td=*][color="#000000"]SLM [/color] [/td][td=*][color="#000000"]353[/color] [/td][td=*][color="#000000"].169[/color] [/td][td=*][color="#000000"]1B[/color] [/td][td=*][color="#000000"].152[/color] [/td][td=*] [/td][td=*][color="#000000"]Vladimir_Frias[/color] [/td][td=*][color="#000000"]23[/color] [/td][td=*][color="#000000"]GRN [/color] [/td][td=*][color="#000000"]333[/color] [/td][td=*][color="#000000"].136[/color] [/td][td=*][color="#000000"]2B[/color] [/td][td=*][color="#000000"].140[/color] [/td][td=*] [/td][td=*][color="#000000"]Chia-Chu_Chen[/color] [/td][td=*][color="#000000"]21[/color] [/td][td=*][color="#000000"]LOW [/color] [/td][td=*][color="#000000"]86[/color] [/td][td=*][color="#000000"].101[/color] [/td][td=*][color="#000000"]C[/color] [/td][td=*][color="#000000"].118[/color] [/td]

Rnk	Name	Age	Tm	PA*	TAv	Pos	Adj
7	Ryan_Kalish	22	BOS	452	.293	CF	.299
19	Anthony_Rizzo	20	PME	626	.302	1B	.285
28	Ryan_Lavarnway	22	PME	550	.264	C	.281
35	Lars_Anderson	22	BOS	548	.295	1B	.278
57	Oscar_Tejeda	20	SLM	560	.268	2B	.272
72	Yamaico_Navarro	22	BOS	462	.254	SS	.268
73	Josh_Reddick	23	BOS	515	.262	CF	.268
83	Che-Hsuan_Lin	21	PME	555	.261	CF	.267
	Reymond_Fuentes	19	GRN	420	.256	CF	.262
	Jose_Vinicio	16	RSX	172	.244	SS	.258
	Jose_Iglesias	20	PME	284	.243	SS	.257
	Will_Middlebrooks	21	SLM	485	.256	3B	.255
	Luis_Exposito	23	PME	536	.238	C	.255
	Peter_Hissey	20	SLM	511	.245	CF	.251
	Michael_Almanzar	19	GRN	542	.251	3B	.250
	Keury_De_La_Cruz	18	RSX	224	.244	CF	.250
	Jose_Garcia	19	GRN	220	.231	SS	.245
	Nate_Spears	25	PME	605	.241	2B	.245
	Daniel_Butler	23	PAW	387	.228	C	.245
	Chih-Hsien_Chiang	22	PME	480	.246	RF	.241
	Bubba_Bell	27	PAW	392	.245	RF	.240
	Christian_Vazquez	19	GRN	303	.218	C	.235
	Alex_Hassan	22	PAW	412	.243	LF	.235
	Gustavo_Molina	28	BOS	126	.215	C	.232
	Dusty_Brown	28	BOS	288	.215	C	.232
	Ray_Chang	26	PME	497	.232	3B	.231
	Felix_Sanchez	20	GRN	254	.222	CF	.228
	Ryan_Dent	21	SLM	427	.213	SS	.227
	Mitch_Dening	21	SLM	508	.230	RF	.225
	Brandon_Jacobs	19	LOW	267	.231	LF	.223
	Kolbrin_Vitek	21	GRN	287	.223	3B	.222
	Tim_Federowicz	22	SLM	456	.204	C	.221
	Derrik_Gibson	20	GRN	567	.206	SS	.220
	Henry_Ramos	18	LOW	184	.224	RF	.219
	Mark_Wagner	26	PAW	151	.202	C	.219
	Gil_Velazquez	30	PAW	259	.204	SS	.218
	Aaron_Bates	26	PAW	489	.224	LF	.216
	Jeremy_Hazelbaker	22	GRN	511	.220	RF	.215
	Sean_Killeen	23	GRN	123	.195	C	.212
	Jason_Thompson	19	RSX	193	.207	2B	.211
	Heiker_Menses	18	RSX	227	.210	3B	.209
	Miles_Head	19	LOW	273	.220	1B	.203
	Jorge_Padron	23	PME	327	.206	LF	.198
	Wilfred_Pichardo	20	GRN	242	.202	RF	.197
	Ronald_Bermudez	22	SLM	480	.204	LF	.196
	Ryan_Khoury	26	PAW	317	.196	3B	.195
	Matt_Sheely	23	PAW	270	.196	RF	.191
	David_Renfroe	19	LOW	227	.191	3B	.190
	Kenneth_Roque	20	GRN	188	.186	2B	.190
	Joantoni_Garcia	19	GRN	236	.184	2B	.188
	Reynaldo_Rodriguez	24	GRN	328	.204	1B	.187
	Jorge_Jimenez	25	PAW	370	.187	3B	.186
	Luke_Yoder	22	RSX	196	.192	LF	.184
	Seth_Schwindenhamm	18	LOW	123	.191	LF	.183
	Moko_Moanaroa	20	RSX	122	.190	LF	.182
	Shannon_Wilkerson	21	GRN	294	.189	LF	.181
	Jason_Place	22	PME	136	.185	RF	.180
	Michael_Thomas	21	GRN	104	.161	C	.178
	Luis_Segovia	23	PME	136	.162	SS	.176
	Bryce_Brentz	21	LOW	288	.177	RF	.172
	James_Kang	22	PAW	174	.168	2B	.172
	Zach_Gentile	23	GRN	381

[Buzzkill Pauley]

As much crap as Lars seems to get around here for Rizzo having "overtaken" him, it's fascinating to see just how equivalent the two players seem to be as hitters.

Not much difference between #19 and #35 overall, when you consider just how many hitters are playing in the minor leagues.

[judyb]

As much crap as Lars seems to get around here for Rizzo having "overtaken" him, it's fascinating to see just how equivalent the two players seem to be as hitters.

Not much difference between #19 and #35 overall, when you consider just how many hitters are playing in the minor leagues.

It's been a little disturbing to me how eager some people have seemed to declare Anderson a bust before he's even had a chance to celebrate his 23rd birthday.

[Eric Van]

It's been a little disturbing to me how eager some people have seemed to declare Anderson a bust before he's even had a chance to celebrate his 23rd birthday.

The take on him now seems to be, he'll be a decent MLB hitter but nothing special.

But the Davenport Peak Translations have him as the 8th best hitting prospect in the minors (and Rizzo as the 7th). How high do you have to rank to project as a middle-of-the-order bat? Figure that there are 60 of them in MLB at a time (average clubs are short a bat and bad clubs are short two) and that the average such hitter has six years of being that good. So that means that 10 such bats come out of the minors each year. Guys on the list I posted earlier pretty much comprise two crop years worth of prospects, so the top 20 guys in the minors project as middle-of-the-order bats. (And just about all the guys on that list have that reputation.)

And the swing he put on Fister tonight for the opposite field double was just beautiful.

Edited by Eric Van, 14 September 2010 - 03:47 AM.

[Hairps]

The question of properly adjusting for age relative for ml players to level is one of the thorniest in sabermetrics. How much air should we take out of Jeremy Hazelbaker's numbers? How much pizazz should we give to Reymond Fuentes instead?

Clay Davenport's Peak Translations are the only comprehensive attempt that I know of to answer those questions. Of course, they should not be taken as gospel, but they should absolutely be used to inform our prospect ratings.

His raw numbers are available at Baseball Prospectus (under Minor League True Average) and you can find stats for the NYP and GCL, even though they are not linked to on the page, by hacking the URL (you can also put in a team code such as BOS to get all the stats for that team).

I'm just getting caught up on this thread. Cool stuff.

EV, do you have the age adjustments pulled out and isolated from the overall translations? If so, I'd be interest in seeing what they are.

[Koufax]

The take on him now seems to be, he'll be a decent MLB hitter but nothing special.

But the Davenport Peak Translations have him as the 8th best hitting prospect in the minors (and Rizzo as the 7th). How high do you have to rank to project as a middle-of-the-order bat? Figure that there are 60 of them in MLB at a time (average clubs are short a bat and bad clubs are short two) and that the average such hitter has six years of being that good. So that means that 10 such bats come out of the minors each year. Guys on the list I posted earlier pretty much comprise two crop years worth of prospects, so the top 20 guys in the minors project as middle-of-the-order bats. (And just about all the guys on that list have that reputation.)

And the swing he put on Fister tonight for the opposite field double was just beautiful.

This analysis seems to ignore your own adjustments, which place Rizzo at #17 and Andersen at #35. If 10 middle-of-the-order bats come out of the minors each year, where does #35 put you? I have seen very little of Andersen myself, and missed last night's game. But your numbers seem to suggest that Andersen is not destined to be a middle of the order bat.

[BLumbergh]

This analysis seems to ignore your own adjustments, which place Rizzo at #17 and Andersen at #35. If 10 middle-of-the-order bats come out of the minors each year, where does #35 put you? I have seen very little of Andersen myself, and missed last night's game. But your numbers seem to suggest that Andersen is not destined to be a middle of the order bat.

Isn't the point that Anderson and Rizzo are the 7th and 8th best hitters -- which is what determines whether one is a middle of the order bat -- but inferior prospects when you take into account the fact that a middle of the order first baseman might be less valuable than a centerfielder who hits in the 6 hole?

[Eric Van]

Isn't the point that Anderson and Rizzo are the 7th and 8th best hitters -- which is what determines whether one is a middle of the order bat -- but inferior prospects when you take into account the fact that a middle of the order first baseman might be less valuable than a centerfielder who hits in the 6 hole?

That's it. The top 20 hitter list, where tthey ranked 7 & 8, was not position-adjusted, the list where they ranked 19 and 35 was. Essentially, the big majority of middle-of-the-order hitters are 1B, DH, and corner OF.

With the position adjustments, there are a lot of C (Wil Myers, Wilin Rosario, Devin Mesoraco, Hang Conger, Salvador Perez, Austin Romine, Lavarnway, Gary Sanchez),, CF (Gregg Halman, Mike Trout, Brett Jackson, Kirk Niewenhuis, Anthony Gose, Daniel Fields), 3B (Alex Liddi, Matt Dominguez, Lonnie Chisenhall, Juan Francisco, Dayan Viciedo), SS (Wilmer Flores, Junior Lake, Eduardo Nunez, Tim Beckham, Jose Pirela) and one 2B (Brett Lawrie) that pass Lars as prospects without projecting to be better hitters. Note that some of these guys may move to easier positions: Myers as a RF and Flores as a 3B still rank ahead of Lars, but Lawrie as a RF, Lavarnway as a DH, Lake or Beckham as 3B, or (obviously) Francisco or Vicideo as 1B do not. So Lars' true prospect ranking is higher than 35 -- but lacking precognitive powers I'm basing this on current positions.

[Koufax]

Got it. That went over my head, but looking back at the thread I see your point.

Having three top hitting prospects (including Kalish)is FAR more hopeful than I thought our system had to offer.

[jon abbey]

With the position adjustments, there are a lot of C (Wil Myers, Wilin Rosario, Devin Mesoraco, Hang Conger, Salvador Perez, Austin Romine, Lavarnway, Gary Sanchez),, CF (Gregg Halman, Mike Trout, Brett Jackson, Kirk Niewenhuis, Anthony Gose, Daniel Fields), 3B (Alex Liddi, Matt Dominguez, Lonnie Chisenhall, Juan Francisco, Dayan Viciedo), SS (Wilmer Flores, Junior Lake, Eduardo Nunez, Tim Beckham, Jose Pirela) and one 2B (Brett Lawrie) that pass Lars as prospects without projecting to be better hitters.

Eric, I think I know the answer, but none of this factors in home park, does it? Romine's home OPS in a noted pitcher's park was .604, but on the road, it was .843, so just curious.

[Eric Van]

Eric, I think I know the answer, but none of this factors in home park, does it? Romine's home OPS in a noted pitcher's park was .604, but on the road, it was .843, so just curious.

All of Clay's translations attempt the best park adjustment he can manage. If he has data to separately adjust the hitting components, he would do that, but at the very least he is taking the run environment and doing a generic translation onto the components.

According to these translations Romine is the 5th best catching prospect in MLB of guys who are considered likely to remain there. And Staten Island C Gary Sanchez was 6th and was the highest rated prospect in all of short-season ball (and Jose Vinicio of the GCL Sox was the highest-rated prospect in all of rookie ball. In fact, the Sox and Yankee organizations dominate in terms of number of guys in the top 100 or more).

It should be noted, though, that as you go further down in the minors you see more and more skill position players (especially C, CF, SS) and fewer and fewer 1B and corner OF. Some of that is because guys have not been moved yet to the easier positions, but that might not be the whole reason.

[finnVT]

I've now got a spreadsheet with every player in the minors. And I've discovered that, while my instincts were correct in thinking that the published Davenport Peak Transalstions were too tough on players in the lower minors, my method of correcting for this (assuming that the prospects at every level are equally good) actually over-corrected. If you take the published translations verbatim, players that were promoted a level tended to gain, on average, .006 of Peak Projection, which is clearly not right. With my fix, though, they tended to lose -.009. Taking 60% of Clay's figure and 40% of my figure yields a Peak Projected TAv which doesn't change when you get promoted, which is as it should be.

I'm just catching up on this, but one thought. It seems like the problem with your initial scaling could be that it doesn't take into effect that at each level, the bottom end of the player talent/skill distribution gets lopped off (since they don't make it to the next level). So you don't want the A average to match the AA average. What you want is that the average of AA should match the average of the top X% of A, where X is whatever % of A players eventually make it to AA. I suspect your 60/40 average is approximating this, but this might give you a slightly more accurate (or at least more easily defended) approach.

[Eric Van]

I'm just catching up on this, but one thought. It seems like the problem with your initial scaling could be that it doesn't take into effect that at each level, the bottom end of the player talent/skill distribution gets lopped off (since they don't make it to the next level). So you don't want the A average to match the AA average. What you want is that the average of AA should match the average of the top X% of A, where X is whatever % of A players eventually make it to AA. I suspect your 60/40 average is approximating this, but this might give you a slightly more accurate (or at least more easily defended) approach.

That's exactly the problem with the initial scaling.

I think the most accurate simple way to fix it is to make sure that, on average, promoted players do not change their prospect status, and the 60 % original Clay, 40% fully scaled mix accomplishes that. I bet it's equivalent to your idea

One thing I didn't look at is whether Clay's mis-adjustment was linear or not. IOW, maybe that 60/40 mix should vary by level. But there almost certainly wouldn't be enough guys at each level to make a definite decision. With three or five years of data, that would be worth looking at.