Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Transaction Oracle > Discussion
Transaction Oracle
— A Timely Look at Transactions as They Happen

Thursday, May 26, 2005

ZiPS In-Season Projections

I’ve added a simple spreadsheet I use to calculate ZiPS projections in-season.  You enter the projection and the season-to-date stats and the spreadsheet automatically calculates the weights and makes a new baseline projection.  The spreadsheet then calculates the rest of the season and spits out the totals. 

Anyone who downloaded my too-soon posting last night should download it again.  I had more time this morning and ironed out what I hope are the last of the kinks (I did a projection for Todd Walker and had a nasty surprise).

Dan Szymborski Posted: May 26, 2005 at 05:23 PM | 41 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Joey B.: posting for the kids of northeast Ohio Posted: May 26, 2005 at 05:58 PM (#1363379)
Just for giggles, I went back and checked out Joe Blanton's ZIPS projections:

Name          ERA   W   L   G  GS     INN    H   ER   BB    K  HR
Blanton      3.89  12   8  30  26   171.0  172   74   39  122  20


Whoopsy-daisy. Better revise that sucker downwards. Way downwards.
   2. Dan Szymborski Posted: May 26, 2005 at 06:07 PM (#1363392)
Guess the K rate caught up with him!

As a side note, I actually uploaded the update file now. This has not been a good couple of days for me!
   3. Dan Szymborski Posted: May 26, 2005 at 06:18 PM (#1363417)
Predicting Blanton the rest of the way:
2005 ZiPS Projection - Joe Blanton
-----------------------------------------------------------
Period       W   L   G  GS   IP    H   ER  HR  BB  SO   ERA
-----------------------------------------------------------
Actual '05   0   5   9   9   44   50   32   8  20  13  6.55  
Rest '05     6   5  18  15   99  100   46  13  25  64  4.18
-----------------------------------------------------------     
Total '05    6  10  27  24  143  150   78  21  45  77  4.91 
   4. Harold can be a fun sponge Posted: May 26, 2005 at 06:20 PM (#1363420)
Dan, why does the hitter projection sheet ask for the projected team, while the pitcher projection doesn't? It seems you'd need them there, too, to back out the park factors in the projection (unless that sheet assumes that the YTD team is the same as projected team).
   5. Dan Szymborski Posted: May 26, 2005 at 06:30 PM (#1363433)
That sheet unfortunately assumes that the YTD team is the same as the projected team. ZiPS Pitcher projections are rather more complex to deal with in this fashion than hitter ones.
   6. Harold can be a fun sponge Posted: May 26, 2005 at 06:35 PM (#1363447)
I updated Dan's spreadsheet so you just type in a player's name and it looks up the projection for you, so you don't have to copy/paste or type it in. Get it here. Dan, feel free to take mine and post it and/or use it going forward.

Now if we could get the YTD stats in a sheet and get a similar lookup working, that'd be awesome. Dan, a request for stuff like this going forward: could you include a standard player ID for each player (retrosheet, Lahman, STATS, DMB, whatever); that would make it really easy for others to add more data, cross-reference, etc.
   7. Hendry's Wad of Cash (UCCF) Posted: May 26, 2005 at 06:39 PM (#1363459)
Vinay, how does this work? Where do I type in the name?

Brian Giles is up when I open the file, but if I type over his name with another, nothing happens but a bunch of errors.
   8. Dan Szymborski Posted: May 26, 2005 at 06:39 PM (#1363461)
That's great Vinay!

How did you do that so easily? That's something I've never played with in Excel - if it's that simple, I can really streamline my projections this year.
   9. Dan Szymborski Posted: May 26, 2005 at 06:41 PM (#1363469)
I've uploaded Vinay's over mine. Anyone else who wants to fix my meanderings is definitely more than welcome to!
   10. Dan Szymborski Posted: May 26, 2005 at 06:43 PM (#1363476)
Oh, the VLOOKUP function. Dumb me.
   11. Harold can be a fun sponge Posted: May 26, 2005 at 06:45 PM (#1363482)
UCCF, you have to type the name exactly as it appears in Dan's sheet (meaning no spaces, no confusing "Joe" vs "Joey", etc). If it doesn't match exactly, you'll get errors.

Dan, Excel's lookup functions are really cool, and once you get the hang of them, they're really easy to use. Just look at the formulas in cells G4:R4 (it's really the same formula, just copied into all the cells).
   12. Hendry's Wad of Cash (UCCF) Posted: May 26, 2005 at 06:47 PM (#1363498)
Ah, thanks. I was putting a space in after the comma, and that was screwing things up.
   13. Dan Szymborski Posted: May 26, 2005 at 06:47 PM (#1363500)
Yeah, I knew about them but thought they were considerably more complex. This will save me a ton of time in the future!
   14. Harold can be a fun sponge Posted: May 26, 2005 at 06:47 PM (#1363502)
There may be something wrong with the pitcher projections. Adam Eaton was initially projected to start 31 games, and has started 10 so far in 46 team games; but he's projected to start only 18 the rest of the way. And I'm seeing similar things for a couple other pitchers I tried.
   15. Dan Szymborski Posted: May 26, 2005 at 06:56 PM (#1363525)
Uh-oh, I think I mucked something new up!
   16. Dan Szymborski Posted: May 26, 2005 at 07:02 PM (#1363541)
I know what I did. HOld on.
   17. bibigon Posted: May 26, 2005 at 07:17 PM (#1363577)
.318/.387/.458 from Renteria the rest of the way.

I'm a bit surprised his awfulness hasn't impacted his projection more.


The only way Renteria is worth his deal is if he matches his ZiPS. Matching his PECOTA won't suffice.

ZiPS gives me hope!
   18. bibigon Posted: May 26, 2005 at 07:29 PM (#1363612)
Why is it asking for team games rather than player games by the way?
   19. Harold can be a fun sponge Posted: May 26, 2005 at 07:38 PM (#1363645)
So it knows how far into the season we are, and how much season is left.
   20. Dan Szymborski Posted: May 26, 2005 at 07:43 PM (#1363662)
But my estimates of playing time got garbled becuase of the framework in place for ZiPS. ZiPS projects whole seasons, not partial seasons, and I missed a few places in which it was using playing time to project future playing time and assuming that partial seasons were full seasons.
   21. Dan Szymborski Posted: May 26, 2005 at 07:48 PM (#1363678)
I've uploaded the newest build (we definitely have to call this a beta), this time with Edgar Renteria and Eric Milton loaded.
   22. Harold can be a fun sponge Posted: May 26, 2005 at 07:55 PM (#1363695)
I'm getting a 404 when I try to download it.
   23. Dan Szymborski Posted: May 26, 2005 at 07:57 PM (#1363703)
Hadn't saved. Should work now - the up-to-date file should be modified 3:55PM.
   24. Harold can be a fun sponge Posted: May 26, 2005 at 08:02 PM (#1363718)
Cool, thanks Dan!
   25. cynic Posted: May 26, 2005 at 08:08 PM (#1363736)
This is great stuff, Dan. Thanks for posting it.

I noticed that Mike Cameron's updated projection is for .259/.372/.483. This seems like a pretty big jump up from his original .244/.338/.449 projection, and it's only based on 65 AB, which sort of strikes me as odd. Am I doing something wrong?

Also, it would be really nice to have in-season stats in the sheet so that everything updates automagically on the projection sheet. That way one could quickly look at updated projections for any player without having to manually update each of the fields.
   26. Dan Szymborski Posted: May 26, 2005 at 08:17 PM (#1363768)
This seems like a pretty big jump up from his original .244/.338/.449 projection, and it's only based on 65 AB, which sort of strikes me as odd. Am I doing something wrong?

Those are 65 unbelievable ABs, though. He can go 0-34 and *still* be beating his projection.
   27. cynic Posted: May 26, 2005 at 09:08 PM (#1363925)
OK, I just wanted to make sure I wasn't misunderstanding how to use the projector.
   28. cynic Posted: May 26, 2005 at 09:10 PM (#1363928)
Also, the line I posted was his projection the rest of the way, not the overall projection (which is even higher). It seemed weird that the projection for the rest of the season would be so strongly influenced by 65 AB, even ones as good as those.

But hey, sounds good to me.
   29. AuntBea Posted: May 26, 2005 at 09:47 PM (#1363990)
The remaining projected ABs for batters is clearly too low. None of the players I have entered have realistic AB totlas for the season. For example, Jeter only ends up with 520 total ABs (instead of a much more correct 610 or so).

How much are you weighting this years results versus the last three years?
   30. Dan Szymborski Posted: May 26, 2005 at 10:33 PM (#1364084)
Are you using the correct file?

Jeter is projected to hit 297/374/446 over the rest of the season to end up at 301/380/438 and 625 at-bats.
   31. bibigon Posted: May 26, 2005 at 11:56 PM (#1364308)
Wow, this hits Beltre pretty hard.

From his great .292/.349/.528 projected line, he drops to a less than great .277/.328/.485...
   32. WillYoung Posted: May 27, 2005 at 07:58 PM (#1366396)
The Twins pen projected out through the rest of the season:

Jesse Crain       10  3 2.86 65  -    66.0  52  21  3  21  57 
Joe Nathan         5  2 2.83 76  -    73.1  56  23  5  24  87 
Juan Rincon        8  5 3.19 68  1    76.1  63  27  3  34  87 
J.C. Romero        2  4 3.34 77  -    70.1  56  26  4  41  61 
Terry Mulholland   1  4 5.17 44  5    87.1  98  50  15  22  34 
Shaggy Guerrier    4  5 4.77 31  12  100.0  106  53  18  29  69 


Those first four are pretty nasty.
   33. WillYoung Posted: May 27, 2005 at 08:07 PM (#1366435)
And the rotation of control:

Silva 8 6 3.66 37 24 187.0 211 76 16 20 70
Radke 12 10 3.94 35 35 231.0 254 101 32 19 136
Santana 14 6 3.26 36 32 210.1 174 76 22 40 256
Mays 7 6 4.77 29 24 151.1 172 80 24 38 55
Lohse 9 10 4.76 31 29 172.0 190 91 24 48 96

165 total walks out of those five.
   34. Cheer and boo and raise a hullabaloo Posted: May 27, 2005 at 08:23 PM (#1366493)
Thanks for this great new tool!

One problem with the hitter projections -- the formula for projected HBP is referencing the cell for SB. So in the spreadsheet I downloaded, D'Angelo Jimenez is projected to get hit 7 times the rest of the season.

Specifically, cell P6 looks wrong. There's no column to the right for HBP, and it looks like there should be one, between AB and AC.
   35. Chris Dial Posted: May 27, 2005 at 08:49 PM (#1366571)
Thanks for the Beltre line, bibigon.

803 OPS+ and that has to be mashing from here on out.
   36. Dan Szymborski Posted: May 27, 2005 at 09:43 PM (#1366686)

Specifically, cell P6 looks wrong. There's no column to the right for HBP, and it looks like there should be one, between AB and AC.


Squashed that bug now. I had fixed the SB and CS bugs earlier, but forgot to do the HBP one. Essentially, that blue area is park neutral information and since there's no HBP park effects, I didn't make a HBP column over that, which made the columns not line up perfectly.

Part of the problems is that I didn't create an entirely new spreadsheet for this but modified the simple ZiPS spreadsheet (that doesn't incorporate multiple teams in one season). This saved me time on the one hand, but also served to cause some adaptability issues.
   37. Most Favored Haitian Status Posted: May 27, 2005 at 10:48 PM (#1366820)
So in the spreadsheet I downloaded, D'Angelo Jimenez is projected to get hit 7 times the rest of the season.

Get me near him and we'll talk.
   38. Cheer and boo and raise a hullabaloo Posted: June 01, 2005 at 04:07 PM (#1374164)
Either I don't understand how you're projecting decisions for pitchers, or something is wrong. I'm looking at cell C6 on the Pitcher Projector, and it's some sort of weighted average. But one of the components is the pitcher's Year-to-Date Decisions. This seems to make the final projected decisions too low.

The example that is in the spreadsheet I downloaded is Eric Milton. He's 3-5, which prorates to 11-18 for the year. His original projection was 8-12. So we have 29 prorated decisions, and 20 decisions from the original projection. I would expect his new projection to be somewhere between those two. But his new projection is 20 decisions, because his 8 year-to-date decisions (not prorated) are heavily weighted in the new projection.

Maybe that's not the best example; it becomes more apparent with, say, Randy Johnson. His new projection is for less decisions than either his prorated YTD or his original projection. This seems wrong, but feel free to tell me that I just don't understand how this is working if that's the case.
   39. Dan Szymborski Posted: June 01, 2005 at 04:49 PM (#1374235)
C+B - good catch. When I fixed the weights to more accurately reflect the seasonal playing time, I didn't fix the decision. I fixed that and now Randy goes 12-5 instead of 10-5 the rest of the way. The updated spreadsheet should be up by about 2.
   40. Cheer and boo and raise a hullabaloo Posted: June 02, 2005 at 06:15 PM (#1377020)
When doing hitter projections for the Angels, there's a mismatch between the "Hitter Projections" sheet and the "Hitter Projector" sheet, so everything craps out and gives errors. The park effects on the Hitter Projector still say ANA, but the Hitter Projections sheet now says LAA.

Hope you don't mind me finding these errors -- I really am using this tool! :-)
   41. Dan Szymborski Posted: June 03, 2005 at 02:59 PM (#1379300)
Good catch - the list of projections was made after the name change while the projection themselves were programmed before.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Infinite Joost (Voxter)
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.3985 seconds
47 querie(s) executed