Their respective stats:
Span
Gwynn Jr.
Gwynn

It's true that Gwynn's rookie performance wasn't anything great, but he did it at age 22, and with equal/better numbers (including an excellent K rate) than Span posted in AA at the same age. And at age 23, Gwynn was posting peak Jason Tyner numbers in MLB -- it's hard to imagine Span coming close to that last year. Age 24 was Gywnn's breakout year in the majors (.351/.410/.444, 23 K's in over 600 AB!), which can't really be on the radar for Span in 2008 (or ever, for that matter).

Meanwhile Gwynn Jr.'s age-23 season in the PCL seems like a very good match for Span's age-23 season in the International League, and they are much closer physical specimens too (taller, thinner). Other than Gwynn Jr.'s walk rate while repeating AA in 2005, their entire careers seems pretty close to me.

I'm not sure if the rest of Span's comparables list would shed much light on the matter... just curious. I know I've mixed up the two Gwynns in conversation before.

I realize that, but if the top 3 are duds, that doesn't bode well for the rest of the list.

Also, I would note that PECOTA does a very good job at predicting future performance on average.

Does the "UPSIDE" score have any direct bearing on the PECOTA projections? I know the comparables do, but the UPSIDE score just seems like an accessory derived from the comparables list.

And in the Gomez case, all evidence suggests their height/weight data is screwy, so I'm not putting much stock in their "average" results for this specific instance. If they happen to be correct, comparing an 18-year-old Dominican's size to the size of 22-year-old ML rookies, it's probably more luck than anything else.

And how the heck is Tony Gwynn the #1 comparable for Denard Span? That would require some serious league/ballpark adjustments, and even then it wouldn't begin to address their disparity in K rates. Are we sure this isn't Tony Gwynn Jr.? He actually seems like a very close match, including height/weight data. It wouldn't explain Span's high UPSIDE score, but who knows, maybe PECOTA got them mixed up too.

They will count as zero, but if you have a bunch of your comparables that contribute nothing, that is going to hurt you compared to someone who has more comparables above average. It certainly penalizes you relative to your competition.

Thus, Gomez' UPSIDE score is going to be the twice the number of runs above average of David Green and Luis Matos, a highly suspect figure based on a highly suspect list of "comparables."

The actual PECOTA forecasts use dozens of comparable players, not just two or three. Also, I would note that PECOTA does a very good job at predicting future performance on average. You can call his list of comparables suspect, but you don't know nearly as many players as the system does, and you haven't compared all of them to Denard Span.

And it sounds like this UPSIDE figure is going to get totally out-of-whack if a superstar like Tony Gwynn shows up on your list -- the way I understand it, Span's UPSIDE figure is going to include "twice the number of runs he [Gwynn] contributed above average." And he won't be penalized for lousy players on his list -- they will just count as zero.

Thus, Gomez' UPSIDE score is going to be the twice the number of runs above average of David Green and Luis Matos, a highly suspect figure based on a highly suspect list of "comparables."

That's a stumper. I wanted to say Steve Sax but I looked him up and his numbers were a bit off.