I've already said this in the dugout, but (IMO) THT's MLEs are not credible - too optimistic, particularly for lower levels.
[That said, its creator is more credible than I am on this subject... still, I'm sticking to my opinion until convinced otherwise.]

At release date, they were not credible. Strasburg aside, the MLEs have been fixed and are in line with what other systems have.

For example, the projection of Jesus Montero, which got a lot of early attention since it's part of the free Yankee sample, initially had him as the best hitting catcher in the game at age 20 and growing into Albert Pujols.

Now he's shown with a 276/318/421 line and growing into a good hitter in his prime.

AROM, I think that they're still too optimistic for non-top prospects and guys at lower levels (this is the chaining v. direct issue ... fixing the "double counting of age" issue improved things quite a bit). I freely admit that as a non-subscriber, looking only at a handful of Yankees is insufficient to draw too strong of a conclusion without taking into consideration my other biases - also, I appear to be a lone wolf in howling about this.
I'm tempted to plunk the $15 to look at this further, but feel understandable conflict at spending money on a product I don't currently believe in.

I'm tempted to plunk the $15 to look at this further, but feel understandable conflict at spending money on a produce I don't currently believe in.

Pay the money! My kids need shoes!

Specifically, what Yankees do you think are too optimistic? Other than Montero I don't think any of their prospects projections make them look like major leaguers in the next 6 years. When it first came out it had Austin Romine as a future 350/500 guy, but now he looks like just a backup catcher.

I really have trouble buying Strasburg as having a sub-3 ERA based solely on college stats in a non-power conference and a few AFL appearances, regardless of what his stuff is.

Let me first reaffirm that my issue is with translation, not projection (where I'm agnostic ... if you can't get the first part right, the second part isn't worth doing).

Going near-alphabetically (my disagreeing on a specific player would be one thing, on a class something else entirely):
2B David Adams split last season between Charleston (A) and Tampa (A+), hitting .286/.373/.443. Minor league splits (which uses similar, but not necessarily identical raw stats because of how they collect them) translates this as .212/.268/.314 (so, essentially big league useless) - THT translates this as .244/.314/.369 (a decent fill in with the bat). I think that the "truth" is somewhere between these two points (as I view minorleaguesplits as too pessimistic). Either way, we've got a pretty big difference in systems - one that ought to be resolved before we place much stock in any results.

Dan, if you're around, how do you translate Adams for last year? What about park/year adjusted DTs (so, leaving their neutralized environment)?

****

Also, I don't think the "direct" idea is totally without merit, not at all - for studs (like Strasburg, though I'm likely to reject his college MLEs) it probably better reflects the rapidity at which his talent will allow him to get to the majors. But using that framework will (imno) result in our assessing the replacement level as higher than it actually is, which has implications for other analysis. I'd like to tentatively explore a weighted average between direct and chaining, based on how highly regarded / experienced / etc...a prospect is, but that carries with it a mess of problems, plus I don't have the time to do any more than pontificate/spout_off at this point.

The minor league splits MLE is something that is a bit outdated. It's based on factors that I sent Jeff Sackmann a few years ago. The methodology has improved since then. I think what that does is for league A, multiply BA by X, etc., where the better way to do it is to break things into components first, multiply K rate by Z and then babip by Y, and recombine into a stat line.

If Jeff wanted my current formula/factors I'd give it to him, but I'm not sure how high updating the mle calculator is on his priority list. I don't really look at MLE's anymore, they are calculated on the fly in the projection process but not output anywhere, so it would take some work to verify what my current MLE is for Adams.

I think Brian's direct method is promising, and I've used it myself to check my defense MLE's. I was thinking it might work especially well on defense, as it seems that defense peaks earlier, and you don't have to worry too much about the assumed improvement from age 21 to 24, for example.

where the better way to do it is to break things into components first, multiply K rate by Z and then babip by Y, and recombine into a stat line.
I strongly prefer this myself, fwiw.

I am interested in the defense MLEs which, admittedly, I haven't looked at yet at all. Correct me if I'm wrong, but isn't there evidence of a short adjustment period that defenders have to go through on the big league level (particularly for OF - call it the upper deck factor)? If so, I wonder if/how this is accounted for.

I think part of my direct/chaining issue is that most people (I think) worry about the Monteros of the world - whereas I'm more focused on how these things handle the Brandon Lairds (see 7).

Other people reading this - please chime in. I think that there are a lot of interesting (to me, at any rate) discussions that can take place in the framework of MLE analysis and design - how you do this says a lot about how you think the game works - yet you rarely see discourse here...

So we're saying that if the Nats were to trade Strasburg, they'd have to settle for less than Doug Flynn and Pat Zachry? :-)

I don't see any egregious projections on the Yankees since Montero is in line with other system and Adams, even if you think his THT projection is a little high, looks like no more than a utility infielder.

But I did find another. Dustin Ackley is being projected as a .320/.400 (that's batting average/OBP) right now. He's only played college plus the fall league. There needs to be much more regression for the college stats. He's a good prospect, but let's not start ordering the HOF plaque just yet.

I thought the Tom Seaver Rule was that if rights to a top prospect are in dispute, the players gets awarded to the bigger market.

and Adams, even if you think his THT projection is a little high, looks like no more than a utility infielder.
You realize that that's much of my point. He may be one day - he isn't yet.

And if the college translations are way too generous, then so are the numbers for recent college grads - like Staten Island's Adam Warren, which Oliver thinks can post a 5.27 ERA for the Yankees right now.

I'm curious about flipping the entire argument. How would someone like mid-2000s AJ Burnett do if transported to college ball? Would he do as well as Strasburg did? Burnett at the time was a pitcher with close to a 96MPH average fastball with the ability to throw it even harder on demand and major league quality secondary pitches. Because his control was below average for a major leaguer with fine but not amazing home run prevention, he was only an 85-90% of average run prevention guy. Against college batters, however, I'd speculate he would walk very few, as he would have no need to try to hit corners as all pitchers must in MLB. An upper 90s fastball anywhere around the plate mixed with MLB-quality breaking balls would probably be unhittable to most college players. I'd guess only the real prospects would stand much of a chance of putting the ball in play with any regularity, and very few college plays could drive the ball. If you buy mid-20s-year-old Burnett could put up the same numbers as Strasburg against college players (I have no idea if this is the case, but I think it's possible), it means a projection has to take into account that Strasburg may not be better than Burnett.

I have Adams at 222/292/329 for 2009, not surprisingly in the middle.

I really have trouble buying Strasburg as having a sub-3 ERA based solely on college stats in a non-power conference and a few AFL appearances, regardless of what his stuff is.

Yeah, sort of. I have trouble seeing him as a mortal lock multi Cy winner. Who was the last pitcher considered this
"can't miss"? Prior?

Who was the last pitcher considered this"can't miss"? Prior?

Ben McDonald says hi

Who was the last pitcher considered this "can't miss"? Prior?

Todd Van Poppel. Was rated so high, A's wouldn't have traded him away for Phil Plainter

Number 1 Overall Draft Picks -- Pitchers


<u> Year Player            Pos Team         </u>
2009 Stephen Strasburg rhp Washington
2007 David Price       lhp Tampa Bay
2006 Luke Hochevar     rhp Kansas City
2002 Bryan Bullington  rhp Pittsburgh
1997 Matt Anderson       p Detroit
1996 Kris Benson         p Pittsburgh
1994 Paul Wilson         p New York (N)
1991 Brien Taylor        p New York (A)
1989 Ben McDonald        p Baltimore
1988 Andy Benes          p San Diego


Source: mlb.com

Thanks Dan - that's about where I think that season belongs. That and a buck will get you a Coke, but...
Granted, you could argue that's not enough to freak out over (and, indeed, the initial wildly inflated numbers have triggered alot of my feelings here), but it's enough not to pay for.

So we're saying that if the Nats were to trade Strasburg, they'd have to settle for less than Doug Flynn and Pat Zachry? :-)

It's too soon, Walt. Too soon for Seaver trade jokes.

I'll let you know when.

Going back further you had Belcher, Moore, Bannister and Clyde as #1's.

[22]

1983 Tim Belcher
1981 Mike Moore
1976 Floyd Bannister
1973 David Clyde

2009 Stephen Strasburg rhp Washington
2007 David Price lhp Tampa Bay
2006 Luke Hochevar rhp Kansas City
2002 Bryan Bullington rhp Pittsburgh
1997 Matt Anderson p Detroit
1996 Kris Benson p Pittsburgh
1994 Paul Wilson p New York (N)
1991 Brien Taylor p New York (A)
1989 Ben McDonald p Baltimore
1988 Andy Benes p San Diego

Prior would've been #1 in 2001 if Minnesota hadn't picked some guy named Mauer....