Mega-Projections

by NYRoyal on Feb 18, 2008 2:25 AM EST

I'm not a huge fan of the various projection systems, but I think they all provide some insight in addition to everything else we know about players. And of course different projection systems often disagree with each other. So I thought it would be interesting to see what the average projection is for Royals players. I took the projections from PECOTA, ZiPS, Bill James, CHONE and Marcel and averaged them. These are the results:

Player AVG/OBP/SLG (OPS) (High-Low)
DeJesus 278/355/410 (765) (786-743)
Teahen 280/352/434 (786) (812-772)
Gordon 267/340/450 (790) (820-757)
Butler 290/358/462 (820) (838-799)
Guillen 275/336/449 (785) (804-771)
Grudz 283/324/392 (716) (734-692)
Pena 264/292/357 (649) (682-632)
Buck 240/309/411 (720) (742-695)
German 281/363/390 (753) (782-739)
Gload 292/335/443 (778) (811-753)
Shealy 260/325/425 (750) (777-722)
Gathright 277/352/347 (699) (712-672)
Callaspo 274/330/383 (713) (744-679)
Olivo 240/273/397 (670) (689-660)
TOTAL 272/333/414 (747)
2007 261/322/388 (710)

Pitcher ERA (High-Low)
Meche 4.42 (4.55-4.26)
Bannister 4.52 (5.19-3.85)
Greinke 4.42 (4.27-4.76)
Tomko 4.73 (4.34-5.20) (some of these were done before the Royals signed him, so they projected him as a Padre)
De La Rosa 5.56 (5.63-5.38)
Davies 5.74 (6.41-5.43)
Bale 4.43 (4.21-4.60)
Soria 3.22 (3.51-2.83)
Mahay 4.46 (4.70-4.22)
Gobble 4.44 (5.09-3.80)
Yabuta 4.78 (4.75-4.81)
Nunez 4.67 (5.17-4.25)
Peralta 4.13 (4.26-3.79)
Musser 4.87 (5.56-3.99)
Hochevar 5.32 (5.98-4.28)
TOTAL 4.58
2007 4.48

For calculating the totals, I used my own predictions for innings pitched and plate apperances. The totals for both hitters and pitchers include an allowance not listed above for a pool of other players who will get some plate appearances and innings. The above hitting and pitching performances would approximately give rise to the following:

Runs Scored: 743
Runs Allowed: 795
Record: 76-86

0 recs | Comment 24 comments

Story-email Email Printer Print

Comments

Thanks for doing this.

76 wins seems like a reasonable projection. I still will project 80, but I am a fan.

My candidates to significantly beat thier projections:
DeJesus
Teahen
Callaspo
Grenkie

Rock Chalk Talk for pretty good KU College Baseball Coverage

by James Quinn on Feb 17, 2008 11:38 AM EST 0 recs

76 wins

A Pythagorean win expectancy of 76 wins means that we're only 2 wins better than last year's Pythag of 74 wins. I really don't think we're only two wins better. I think Meche, Bannister and Greinke will all be better than the above projections and given the innings they will pitch, that's going to be huge. And I won't go player-by-player, but I think many position players will beat their prediction by a little or a lot.

RS: 765
RA: 755
Record: 82-80

I probably disagree with you.

by NYRoyal on Feb 17, 2008 11:56 AM EST 0 recs

2007 pythag

The thing about the 2007 pythag expected wins is it's kind of distorted by the three 17 run games. Obviously they happened, so the Royals should get credit for them, but if you were to throw those games out, I think you'd lose about 5 wins. I just think that given the Royal's shaky offense last year, these games mostly make the win expectation artificially high.

by PopeSoria on Feb 17, 2008 12:42 PM EST up 0 recs

That is true of every team, every year

Every team has some outlier games where they scored a hell of a lot of runs. For each of those teams, if you took out those high-run outliers, it would change their expectancy by a handful of wins. There's a reason that you count those runs...because they were scored.

You should also realize that the 2007 pythag expected wins is distorted by some games where we allowed a lot of runs (including the three games were we gave up 17, 16 and 13 runs). If we take them out, the run expectancy changes significantly too. If you take out outliers on one end, you have to take out the outliers on the other end too. But there's a reason you don't take out the outliers. They are all part of the bigger picture.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 12:52 PM EST up 0 recs

Opponents

There were also games when the Royals allowed a ton of runs. In the past, that would have been commonplace, but last year it made for a bit of an anomaly when it occurred.

If I had a rocket ship, I'd never want to land.

by dman126 on Feb 17, 2008 6:31 PM EST up 0 recs

we should keep these handy for awhile

by royalsreview on Feb 17, 2008 1:53 PM EST 0 recs

If our pitching is that bad

I'll eat my favorite Royals hat.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 2:03 PM EST up 0 recs

I'm actually sort of encouraged

We could potentially have 5 starters with ERAs in the mid to low 4's (Greinke, Tomko, Meche, Bannister, Nunez/Bale).

I think Greinke could end up as high as 3.20, 175 K's, so if the rest of the rotation holds, we're looking at perhaps the best 1-5 in the AL Central.

This space intentionally left blank.

by marbotty on Feb 17, 2008 2:20 PM EST up 0 recs

One clarification

All of the projections for both Nunez and Bale have them as relievers, not starters (0 GS). I'm sure if they were projected as starters, the ERA's would go up.

I think the top 3 in the rotation will do well (Greinke 4.00, Meche, 4.00, Bannister 4.25) and then we'll get something serviceable from the #4 and who knows from #5. That's a pretty good rotation.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 2:26 PM EST up 0 recs

weird yabuta projection

They all have him within .06 ERA?

Considering how hard it is to gauge Japanese talent, I'm amazed they all basically came to the same figure.

This space intentionally left blank.

by marbotty on Feb 17, 2008 2:06 PM EST 0 recs

Yabuta

Only PECOTA and DiPS had any projection for him at all.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 2:13 PM EST up 0 recs

Good work

I've been wanting to do something like this but was too lazy! I've skimmed over the projections and overall I didn't really like what I saw. My optimistic side just noted that Bill James seems to see us doing the best.

I don't get the general consensus of German being a 753 guy, I think his OBP will be 370-390 alone. Just give him the bat and we'll eat it somewhere in the field! Who knows. I guess the only thing I really agree with is that they already see Butler as our best hitter. Good work on this, and I also believe we will win more games than the pythag expected. Thats why we're fans!

Alex Gordon in '08

by RoyalJHWKR on Feb 17, 2008 3:45 PM EST 0 recs

German

My person prediction for German is .285/.370/.400 (.765), so I don't think the above projections are far off. German is somewhat hard to project because he didn't have significant playing time until 2006. So basically we just have two seasons of performance to evaluate him from.

2006 .326/.422/.459 (.881) 331 PA
2007 .264/.351/.376 (.727) 405 PA

I think the real German is somewhere in between those two performances. I just don't see him as being able to maintain a batting average well over .300. And without making that kind of effective contact, he'll have a tough time getting his OBP up over .370. And he doesn't have a lot of power. But he could easily surprise me and put up numbers better than this.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 4:59 PM EST up 0 recs

while not foolproof

i know he's getting old for "breakout" seasons, germans minor league numbers are what makes me think he would beat these projections, especially if he gets consistent AB's. i despise the whole "he'll get worn down" theory. maybe a 794 career ops doesn't translate to a 794 mlb ops, a 770-780 isn't' unreasonable as a hitter progresses.

Alex Gordon in '08

by RoyalJHWKR on Feb 17, 2008 5:18 PM EST up 0 recs

Nice job

ZiPS, not DiPS, right?

I am not sure about using Marcel, which is more of a regression analysis than actual projection system. Its creator, Tom Tango, even jokes that he does not stand behind it. Marcel is named after the monkey (he must have been a Friends fan back in the day) with the idea that the analysis is so easy that even a monkey could do it. It simply takes a weighted average of the last three major league seasons (weighted 5-4-3) and adjusts for age. It uses only major league numbers, so the "projections" for first and second year players (and the Royals have a lot of them) really should be discounted or thrown out altogether.

by Gopherballs on Feb 17, 2008 4:58 PM EST 0 recs

Whoops, that's right.

It is ZiPS. I should edit that.

And you are probably right about Marcel. But they wouldn't change the overall average much.

I probably disagree with you.

by NYRoyal on Feb 17, 2008 5:01 PM EST up 0 recs

MARCEL

actually works pretty well for established major league player, ie players for which there are at least 3 years of major league data.

It's problem is young players.

visiting A's and Mets fan

by rfloh on Feb 18, 2008 1:27 AM EST up 0 recs

The Royals roster

is a who's who of players with less than 3 years of service time.

by Gopherballs on Feb 18, 2008 11:07 AM EST up 0 recs

My prediction

Every starter outperforms their OPS projection.

by doublestix on Feb 17, 2008 4:59 PM EST 0 recs

DeJesus

I wonder what it was about 2007 that has projections screaming that it wasn't a fluke for DDJ. He'd had very steady production for 2 and a half years and then had an off-season, but it seems like all the projections are saying that he'll basically repeat his off year.

If I had a rocket ship, I'd never want to land.

by dman126 on Feb 17, 2008 6:29 PM EST 0 recs

I Think It

Might be the significant increase in games played and plate appearances that has some people spooked. Ironically, it was his first year without significant missed playing time because of injury, but I think the reasoning is that he "wears down" if he plays a full season.

I'm not getting older....oh, wait, yes I am....and slower.... and weaker. God, this is great!

by philofthenorth on Feb 17, 2008 7:20 PM EST up 0 recs

At least for PECOTA

the two top comps for DeJesus are Erstad and Kotsay, who were both better players before age 27 than afterwards (Kotsay's career year at age 28 aside).

by Gopherballs on Feb 18, 2008 11:09 AM EST up 0 recs

Great work NYRoyal

I've downloaded ZIPs onto my Diamond Mind simulator. I've run about fifteen seasons, with the Royals generally winning 74-76 games. They topped out at 78-84 one season, and bottomed out at 61-101 one season.

For some of the sims I plugged Soria in as a starter midway through the year. That did seem to improve things. A few times I put him as a starter the whole year, and he turned in a 19 win performance one season.

I think you're right about those pitching projections to be a bit pessimistic, but that's still encouraging even if they hit those marks.

Relive Royals History at royalsretro.blogspot.com

by RoyalsRetro on Feb 18, 2008 3:10 PM EST 0 recs

Lineup calculator

Using Baseball Musing's Lineup Calculator, and using the above projections for our lineup of:

CF DeJesus
2B Grudzielanek
LF Teahen
RF Guillen
DH Butler
3B Gordon
1B Gload
C Buck
SS Pena

It predicts 4.827 runs/game, which goes to 782 runs for a season, which you can probably reduce somewhat because that doesn't account for bench players.

It also suggests the best lineup would be:

LF Teahen
DH Butler
2B Grudzielanek
RF Guillen
3B Gordon
C Buck
1B Gload
SS Pena
CF DeJesus

Having DJ last seems very odd.

Relive Royals History at royalsretro.blogspot.com

by RoyalsRetro on Feb 18, 2008 3:17 PM EST 0 recs