Can you beat the monkey?
6 Comments Published by Colin Wyers on Thursday, February 14, 2008 at 2:06 PM.
Al Yellon over at Bleed Cubbie Blue has kicked off his 2008 community projections project with Geovany Soto. If you haven't already, I strongly encourage you to go over there and contribute a forecast. I would appreciate it if you would:
Given my concerns over data pollution, I'll hold off on further discussion until later.
- Take the exercise seriously.
- Be honest. And don't be afraid to be honest.
- Don't worry too much about what projection systems, or other people, have to say. What I'm looking for is your opinion.
Given my concerns over data pollution, I'll hold off on further discussion until later.
Labels: Baseball, Chicago Cubs, Projections
It will be interesting to see what the average the bloggers come up with for each player. I was a wee bit liberal with Soto (I believe, anyways) but I think he's going to be the real deal and have a good year and post some decent numbers.
I think this is suitably buried enough that we can talk about the results briefly without worrying about affecting the results.
The reporting on Soto isn't finished yet, so these numbers could shift right now, but the average forecast for Soto worked out to a .278 wBA.
Compare that to the forecast chart for hitters. That's actually lower than what every projection system but PECOTA projected Soto for.
Quite intriguing, given the fact that there seems to be a vocal minority that says Cubs fans (or at least the ones on BCB) are too optimistic about Soto. Seems like the BCB community actually has very reasonable expectations.
Colin, not sure if this should be here but...I found some discrepancies in the spreadsheet.
You might re-average your columns as I come out with quite different numbers for everything except triples.
Also, isn't SLG:
1B+(2Bx2)+(3Bx3)+(HRx4)/AB?
You have total hits and don't have HRx4.
I moved all the items (from the 'average' line) down to line 37 in the spreadsheet and also added an item (1B at N39). Here's what I cam up with:
the formula at N39 is:
=(E37-(F37+G37+J37))
(total hits-2B-3B-HR) which gives me total 1B.
Then I use the following at N37 for SLG:
=(N39+(F37*2)+(G37*3)+(J37*4))/B37
This gives me (using the lines from row 37 that I re-averaged) a SLG of .454 where the spreadsheet shows .524.
No time to go into detail with this, but extra base hits are included in Hits, which is why my formula for SLG looks different.
My averages are weighted by PA. Send me your spreadsheet at pontifexexmachina@hotmail.com and I'll take a look.
Okay, I found (and fixed) the error in calculating the rate stats. The funny part is, wOBA/wBA stayed exactly the same. The upside is, now I'm getting proctological in figuring out how everything was calculated, and so I caught a few other things.
I also figured out something that was bugging me for a while - players whose OPS and wOBA made no sense next to each other. Going back in and recalculating OPS from the counting stats showed some major discrepancies - in the case of your entry, for example, you forgot to add walks to the denomenator for OBP. That was a rather common error, actually.
Appreciate the help and input. Just getting up to speed on all the stats and formulas.