BP's Clay Davenports DTs for Sox Prospects

[jeremyb1]

Clay Davenport's DTs are projections of players future performance based on the major league translation of their current season, the level, and the players age. The projections are affected by playing time and not rate statistics so when looking at a small number of plate appearances, a player will not rank particularly highly.

http://www.baseballprospectus.com/st...s/eqamin2.html

The above link features links for the DTs for pitchers and hitters this season.

The Sox currently do not have any players ranking in the top thirty for pitchers or hitters (Jeff Francis is currently the number one pitcher, David Wright the top hitter). However, a few of our guys particularly in the lower minors rate solidly. Currently a major league projected 21 RAR (runs above replacement) would put a player in the top 30. The Sox top DTs for the season are:

Mike Spidale - 18
Brian Anderson - 15 (9 at WS, 6 at Bir)
Chris Young - 13
Ryan Sweeney - 12 (5th best in the Carolina League)
Casey Rogowski - 10

And two notables no longer in the system:

Jeremy Reed - 16
Michael Morse - 12

For pitchers an RAR of 25 would put a player in the top 30.

Felix Diaz - 24
Brandon McCarthy - 20
Josh Stewart - 19
Matt Smith - 17
Brian McNichol - 16
Arnie Munoz - 15 (19 Birm, -4 Char)
Josh Fields - 14

And one notable no longer in the system:

Jon Rauch - 17


In a lot of ways these should be taken with a grain of salt as more of a fun calculation, the most important caveat being that the calculations are based on on this season. The projections also are not calculated using more sophisticated information such as height and weight. Only, AVG, OBP, SLB, BBs, Ks, SB, and CS are used.

[Randar68]

Originally Posted by jeremyb1

In a lot of ways these should be taken with a grain of salt as more of a fun calculation, the most important caveat being that the calculations are based on on this season. The projections also are not calculated using more sophisticated information such as height and weight. Only, AVG, OBP, SLB, BBs, Ks, SB, and CS are used.

It's fun to try to come up with a way to characterize these things, but some of these White-Sox-only examples basically invalidates his method/theory he applied...

[jeremyb1]

Originally Posted by Randar68

It's fun to try to come up with a way to characterize these things, but some of these White-Sox-only examples basically invalidates his method/theory he applied...

Well in my opinion the obvious problem is that we're less than two thirds through the season and the remainder of the players' minor league history hasn't been taken into account. We're looking at a pretty small sample size here.

[Randar68]

Originally Posted by jeremyb1

Well in my opinion the obvious problem is that we're less than two thirds through the season and the remainder of the players' minor league history hasn't been taken into account. We're looking at a pretty small sample size here.

How do you account for whether the kid came out of HS or college as a jr or sr(different development/age schedules), things like Tommy John that delay a prospect's development but not a long-term indicator of failure anymore?

[jeremyb1]

Originally Posted by Randar68

How do you account for whether the kid came out of HS or college as a jr or sr(different development/age schedules), things like Tommy John that delay a prospect's development but not a long-term indicator of failure anymore?

Well again Tommy John falls into the large problem that we're only looking at one season. PECOTA which won't be around for months, uses three seasons of data so for most prospects, a season where a player is missing significant time tends to lead to comps of other players with injuries that kept them out. But surgery is a good place where you have to use your own analysis and insight and proves why statistical models are never perfect, they're simply an aide.

As far as whether a player came out of HS, JUCO, College, etc. the belief is that aside from age, which is taken into account, those factors matter very little. Obviously a player that doesn't leave college until he's a high school senior will never be able to reach high A at 19 as Sweeney has but theoretically if he's ability's the same then when he's 22 and making his debut he'll be at a high level or will advance extremely quickly. A college player can't ever reach the majors at 19 or 20 but those events are extremely rare. Players like Prior, Alex Fernandez, or Ryan Wagner have made it in little or not time.

[Randar68]

Originally Posted by jeremyb1

As far as whether a player came out of HS, JUCO, College, etc. the belief is that aside from age, which is taken into account, those factors matter very little. Obviously a player that doesn't leave college until he's a high school senior will never be able to reach high A at 19 as Sweeney has but theoretically if he's ability's the same then when he's 22 and making his debut he'll be at a high level or will advance extremely quickly. A college player can't ever reach the majors at 19 or 20 but those events are extremely rare. Players like Prior, Alex Fernandez, or Ryan Wagner have made it in little or not time.

I guess I just don't buy that. For example, the players you mention are the extreme rarity, like Pujols. There is an adjustment period for any player changing levels. A player comes along every 5 years that performs as Jeremy Reed did, but it is too rare an event to base any kind of statistical analysis upon. HS players have to physically and mentally mature in addition to the challenges college hitters face, such as, the transition to wood bats, the more talented pitching, and vice versa for pitchers adjusting to more patient hitters that destroy mistakes and keeping the ball down in the zone.

You see, these are all learned abilities and factors, that are not simple in nature, yet talented players all learn these various factors at different rates/stages.

I guess, IMO, there are simply too many factors when it comes to HS/College age kids developing to come up with any kind of relatively simple statistical formula to predict it.