Khalil Greene Results

by John Sickels on Mar 27, 2005 5:37 PM EST in News Comment 13 comments

Khalil Greene

Community Projection: .279 BA, .352 OBP, .458 SLG, 16 HR, 9 steals in 531 AB (16)
Baseball Forecaster: .266 BA, .325 OBP, .424 SLG, 14 HR, 5 steals in 515 AB (14)
Baseball Prospectus: .260 BA, .327 OBP, .424 SLG, 15 HR, 3 steals in 467 AB (17)
Bill James Handbook: .273 BA, .335 OBP, .446 SLG, 15 HR, 4 steals in 487 AB (16)
ZIPS BTT Projection: .275 BA, .342 OBP, .428 SLG, 15 HR, 3 steals in 528 AB (15)
John Sickels Experi: .281 BA, .347 OBP, .422 SLG, 12 HR, 4 steals in 502 AB (13)

By OPS

Community: .810
Bill James: .781
ZIPS BTT: .770
Sickels Exp: .769
Prospectus: .751
Forecaster: .749

The Community is the most optimistic projection of all, even in the steal department. Do we know something that the other systems are missing, or did we just have a bunch of Greene Fanboys post? I personally believe that Greene WILL post Community-style numbers eventually, but perhaps not in 2005. I'd be surprised if his slugging percentage is that high in `05, given park factors, but I don't expect any huge slippage in his numbers. My own system projects a slight slippage in the SLG department from last year, but a steady OBP and slightly improved batting average. I think his breakout will be in '06. Greene is a solid player, and if personality and work ethic mean anything, he has a chance to get far beyond "solid" eventually.

0 recs | Comment 13 comments

Story-email Email Printer Print

Comments

community projections

community is always on the high end it seems. I think its some self selection.

I also think the community just picks rate stats out of a hat rather producing a counting stats line and seeing what rate stats that would produce. I distinctly recall some projections last week that predicted .280avg .550 slg and 25 homers for an 500+ab player. That player would have to go nuts with doubles and triples to meet that SLG.

Same goes with some Francis projections. Some had him giving up 20-30 less hits than IP but still posting a 4.75 era. Just doesn't add up. When you post Francis' projection... it might be useful to also provide an estimated component ERA derived from the other data provided.

Oklahoma St, Louisville, Duke, UNC. Stinky Bracket. STINKY BRACKET

by natsfan2005 on Mar 27, 2005 6:03 PM EST reply actions 0 recs

proj

The first projections we did often had 50-60 entries. Lately there have only been 20-30, and the numbers look wonkier.

by John Sickels on Mar 27, 2005 6:10 PM EST reply actions 0 recs

traffic

how is overall site traffic doing?

Oklahoma St, Louisville, Duke, UNC. Stinky Bracket. STINKY BRACKET

by natsfan2005 on Mar 27, 2005 6:18 PM EST up reply actions 0 recs

oop

That should be "weirder" not wonkier. But natsfans' point is well-taken.

by John Sickels on Mar 27, 2005 6:10 PM EST reply actions 0 recs

math on Greene's SLG

If my math is right, Khalil will need 47 doubles with 0 triples to meet that SLG. Obviously you can subtract 2 doubles for every triple added.

Lets' say 4 triples:
Will a .279 hitting sophomore SS playing in a tough hitters park hit 39 doubles, 4 triples and 16 homers? Thats a lot of extra base hits to come out of 148 hits. Not impossible, just not likely. I just don't think a community projection would come out that high if more people calculated rate stats from components rather than fabricating the # from intuition.

Oklahoma St, Louisville, Duke, UNC. Stinky Bracket. STINKY BRACKET

by natsfan2005 on Mar 27, 2005 6:16 PM EST reply actions 0 recs

agree

I agree with natsfan on this. I think people are just a bit too lazy to break out the mathematics required to calculate the actual SLG/OBA.

John, maybe you should just have people predict doubles, triples, homers, etc. and then calculate from there. It's a bit more work for everyone since they'll have to think a bit instead of just throwing any old numbers down in the comment box, but maybe that will weed out some of the wonkier/weirder numbers people are submitting.

by sabernar on Mar 27, 2005 6:22 PM EST reply actions 0 recs

Isn't the point to make numbers up?

I thought the reason we were doing this was to see how educated fans' guesses fared in comparison to formalized prediction generators like Pecota... I know some people have been using their own projection programs, but that's their prerogative; some people just don't have the time, knowlege, or inclination to devise such a thing. I've personally been looking at past performance record, making some basic assumptions about park factors, age, etc., and making up a stat line that "looks right." It's obvious that some people haven't thought their numbers through as much, and perhaps a little more care would make it more accurate, but the system is what it is: an aggregate of people's guesses. And if someone guesses that Khalil Greene will slug 550 in Petco, well, that's their guess, and the system ought to include it.

Or am I totally off-base with this?

by delomir on Mar 27, 2005 6:35 PM EST reply actions 0 recs

guessing

Yes, I think we are asking for guesses.

However, i think people's guesses on rate stats aren't in line with what they would their guesses would be on the components counting stats. I don't expect people to develop rigourous statistical models like John or BP, they can and should guess. But our community projections could be more accurate if people took the 30 seconds to calculate what their guess on AVG, 2B, 3B, and HRs would produce for a SLG. We're already on computers - shouldn't be hard to open the calculator program.

Either way, people can do whatever they want. I was just offering an explanation for why our OPS projections nearly always are the highest.

Oklahoma St, Louisville, Duke, UNC. Stinky Bracket. STINKY BRACKET

by natsfan2005 on Mar 27, 2005 6:44 PM EST up reply actions 0 recs

Rates v. Components

I suppose people might be more comfortable with just the counting stats. I actually found it easier to come up with the rates first, and estimate counting stats from that. Though I may try doing the counting stats first next time around and seeing how that turns out... I don't know, that just seems backwards to me :)

I like the idea of including a component ERA with pitchers with the community predictions. Maybe some people predict a pitcher will be just plain lucky/unlucky in the ERA/record departments, or are worried about the team's defense, etc. Perhaps John could do that for the hitters' rate stats too, if those are more out of line than the counting stats. I'd like to keep predicting the rates, though. I doubt I'm the only person who goes rates-->totals, not totals-->rates.

by delomir on Mar 27, 2005 10:15 PM EST up reply actions 0 recs

projection

Everyone has their own "neural net" model in their head. Personally I think the better way to do it is to break down the component stats before you get the percenage...2B, 3B, HR to get SLG, as natsfan and sabernar suggest. But the point is to get as many "model outputs" from people's heads as possible, and see how the aggregate compares with what other systems show and what reality itself will show.

Now, there will be some weird projections that could throw things off. But the same thing happens in weather computer models. So, basically, I want people to come up with their projections however they want to. The only caveat is that you have to take it seriously. Now, what "seriously" means will vary from person to person. Some people no doubt have complex spread sheets, and some may just be slopping numbers down. But so far there have only been a handful of entries that look outside the realm of possiblity.

by John Sickels on Mar 27, 2005 6:44 PM EST reply actions 0 recs

Greene

I am optimistic about Greene and - based on some minor numbers - kept his extra base hits roughly the same as they were last year but projected then through 155 games. Last year, he split time early. His slugging last year was .446 and my projection had him at .448. That gave him 17 HR, 38 doubles, and 3 triples with a .282 BA. I actually don't think that is too out of line. Petco has a lot of outfield space and I think an extraordinary amount of doubles is not unreasonable.

I do agree though that the community is usually too high and I admit that I could be too optimistic about Greene. But considering that he had 31 doubles and 4 triples last year in 139 games, I don't think it is too unrealistic.

by count sutton on Mar 28, 2005 9:56 AM EST reply actions 0 recs

greene vs loretta

CountSutton,

Sounds like you put thought into the quantity and distribution of extra base hits Greene would have to hit to produce your SLG estimation. Thats great. Thats my only point. We can be more accurate if we all do that like you have.

Even if someone thought Greene is going to hit 22 homers I wouldn't dispute their right to think that. I was just wondering if when most people picked the combination of ".279 BA .458 SLG 16 HR in 531 AB" they realized that Greene would have to collect basically the same distribution of extra base hits as Mark Loretta did last year. Loretta achieved that distribution by knocking 50 more hits in 100 more atbats than the community projection for Greene and lead major league 2nd baseman in VORP. Perhaps if some people knew their SLG projection translated to that they'd be more conservative.

Oklahoma St, Louisville, Duke, UNC. Stinky Bracket. STINKY BRACKET

by natsfan2005 on Mar 28, 2005 10:56 AM EST up reply actions 0 recs