Lookout Landing

Alexander Calloway - January 9, 2010 reply

Odds are the Angels and Rangers blogs are going to over rank how good their players are

Lone Star Ball is a very good blog

I don’t know if they’ve done a similar chart but my guess is the differences will be in playing time not performance projection.

LSB

A commenter there posted a fanpost on the projection of the Rangers shortly after Matthew posted his first one. Totals are remarkably similar but they admit that they’re very conservative.

ThundaPC - January 9, 2010 reply

exactly

I think if you see Chris Davis as a 0.5 player (who could be anywhere between +4, or a -1), the Rangers are going to go another direction. Same with Teagarden.

That said, this is all a very fair analysis – Anaheim has a lot of low ceiling, low risk guys; Seattle is heavy at top (with Lee and King Felix, Ichiro and Gutierrez), but has a good deal of filler; and Texas is very young and full of high risk but high ceiling guys. I wouldn’t be surprised at any of these teams winning.

JBImaknee - January 11, 2010 reply

I keep up with the Rangers constantly

There’s no way IN HELL Brandon McCarthy is going to be a starter in Texas. There are too many other options that have to fail in order for him to make the rotation. The Texas rotation will likely be:
Harden
Feldman
Hunter
Holland
Feliz

texaschief - January 12, 2010 reply

In saying that are you implying Matthew overrated some M's?

Because being able to project players without even a little bias is tough. I wouldn’t be overly upset even if they did.

njd.aitken - January 9, 2010 reply

I think he's saying just the opposite, that the Angels and Rangers blogs would be more biased.

Mariner John - January 9, 2010 reply

They do good comics

From time to time. So they have that going for them, which is good. The post-print journalism world needs its serious analysis, and it needs its comics pages and horoscopes.

But mostly, yeah, they’re the Go2Guy of blogs.

This is handy, thanks

I actually don’t exepect the picture to change much at this point since the only impact players left out there are Pineiro and.. well that’s about it.

You’ve changed your projections a little since last time – Johnson gets bumped from 0.5 to 1, which is totally reasonable since a catcher needs to really stink to be so worthless, and Guti I think you had at 3.5 before.

I still think your 2.5 WAR projection for the pen is too low, even with reduced playing time. The last few Mariner pens have produced win totals of 4.6 (488IP), 5.6 (505IP), 2.8 (530IP), and 3.1 (490IP). The lowest total for the last four years was 2008’s disaster, and there you have to keep in mind that the closer stunk and that’s where most of the pen’s value comes from.

You mentioned last time that the low projection for the pen is partially due to less playing time on account of the improved pitching staff, but the counter to that would be that the lost innings are taken away from the worst pitchers and it ought to actually improve the pen as a whole. Overall, it looks like you shortchange every team a bit so the overall picture wouldn’t change much perhaps, except to widen the gap over Texas which gets a fair shake.

Speaking of Texas, I’m curious how you are projecting Harden, mostly because I always wonder how injury-prone players ought to be projected. Looks like you’re expecting him to sit out a month or so.

Finally, what I like about the Angels projection is that while you defaulted to allocating playing time to the best players the actual breakdown will likely be working against them. Scioscia seemed to show a commitment to giving Mathis at bats beyond a strict platoon, GMJ will be getting time in the outfield, and apparently Matsui expects to play a little out there too. The team is clearly writing off OF defense, which is going to be more important for them than other teams since their rotation had one of the lowest GB rates in the league and hasn’t changed. In fact, replacing Lackey with Kazmir should make it worse.

With regards to the pen, Matthew is probably factoring Lee into this. His 5.5 WAR is going to take some value from the pen.

EnglishMariner - January 9, 2010 reply

Which after re-reading what you posted in pretty much what you raised anyway, sorry.

EnglishMariner - January 9, 2010 reply

Doesnt he just take away innings from the worst bullpen pitchers though?

League and Aardsma should still pitch about the same?

Seems realistic to me at this point in time.

It’s about how I figure things would come out if the season started tomorrow. I expect a close race with the Angels and Rangers with the Ms just squeaking in ahead of them. However, there is still time for some significant changes to rosters and luck will take a hand as it always does. It is great to be in for the pot with some good cards for a change and going from plan B to plan Z!

maqman - January 9, 2010 reply

were the Mariners just really lucky last year?

looking at their runs scored and runs against, it seems like the answer to this is mostly a strong yes.

But I feel like this team is much improved over last year’s team, but our projected wins have us treading water.

Am I missing something here? or is it just that last year was lucky (i.e. chemistry matters! it’s proof!)

Snuffleupagus - January 9, 2010 reply

They got very unlucky with runners in scoring position. They shouldve scored more runs.

Overall they were just slightly lucky.

Hilariously though, they still did finish dead last in wOBA

seattlebruin - January 9, 2010 reply

Not really no

what you’re missing is that you cannot project players to repeat their performances from the last year. Felix might repeat his ridiculous 09, but it would be foolish to project it. Ditto Ichiro, Guti, Aardsma, Figgins, Lee, etc etc.

Last year’s team was actually quite unlucky in terms of run production because of a league-low and unsustainable OPS with runners on. Where they got lucky, if you can call it luck, is the performances they got out of some players, but if you add up the WAR totals you get 20.5 for position players and 16.5 for pitchers. Add that to the ~48 ‘guaranteed’ wins and you get 85 so they were spot on.

so we're seeing regression in these projections?

not to just throw out the over-used regression? but Ichiro, Guti, Aardsma would be projected lower than their performances last year.

I guess I’m confused because Figgins is that we upgraded at 3b, ss, sp and dh. We downgraded at 1b. So why are we projected to win the same number of games as we did last year? am I missing something, or is it expected regression in the performance of Ichiro, Guti and the bullpen?

Snuffleupagus - January 9, 2010 reply

Yes, regression.

Matthew is listing what he considers the true talent level of the players in question. Guti had a 6 WAR season, Ichiro 5, Figgins 6, Aardsma 2, Lopez nearly 3 and so on. On the pitching side, Felix and Washburn outperformed any reasonable projection as well.

So, while these performances were not lucky (since WAR is context neutral), they were out of line with those players established talent level. Maybe they represent a real step forward for each player, maybe not. But projecting them to repeat is not a good way to go about it. There have been quite a few posts on this subject both here and at USSM. In fact, I think when DC presented his offseason plan he laid out in a much more cogent way what I am trying to articulate here.

I Lick Squirrels - January 9, 2010 reply

makes sense

I guess I was counting that as ‘luck.’ In that having several players perform above their expected value at the same time is a lucky coordination that makes the team look better than it was. I guess I feel like performances ‘out of line with established talent level’ are ‘lucky’ when you have several such performances that occur in the same year on the same team.

makes sense to me.

Snuffleupagus - January 9, 2010 reply

Looking at RS vs. RA isn't always the best way to do things.

Consider the fact that there was a huge difference in performance between the best bullpen arms (Aardsma, Lowe) and the worst (Tits et al) last year. Now, if the game was close, Lowe and Aardsma usually pitched. However, if the M’s were losing by 5 runs or something, since with their offense there wasn’t much of a chance of them coming back, Tits or another mop-up guy came in, and usually allowed a few more runs, pushing back the run differential and therefore the projected pythag record even though the game was basically already decided. Take this game for example. The M’s were already down 6-1 in the 4th, so rather than use the good relievers, they brought in Vargas, Messenger, Silva, and French, giving up another 4 runs in the process even though the game was already basically decided.

Pythag record is good for quick calculations, but it often leaves things out. When you look at last year’s team in terms of WAR, they were right about where they should be.

Yeah, I think pretty much every in-depth analysis of pythag

includes caveats about bullpen quality. It appears to have a much higher impact on team record than the total ‘talent’ it contains, compared to the rotation or lineup.

misterjonez - January 9, 2010 reply

WAR has the same problems as Pythag, in that it doesn't consider timing of events

I would argue that it shouldn't.

Pythag… maybe. A 30 to 1 game does throw off the Pythag quite a bit. But WAR is about the skill level of individual players, and whether or not they hit before or after someone else that is good is irrelevant towards judging that individuals skill level. So if they did start adjusting for timing, it would become a worse system for telling how someone actually performed. A guy who hits a double every at bat but never knocks anyone in or scores a run would have a WAR of zero. Blech.

I'm not arguing for WAR to include timing, simply saying that it doesn't incldue timing

The problem with using WAR is that it isn’t meant to be a team metric. A players WAR is how many wins he would contributed, given neutral timing, if he was put in an AVERAGE lineup. The problem is that teams each have their own run environments, in both the pitching and hitting side of the ball, so the actual values that make up wOBA need to be adjusted for that.

BaseRuns is the best way to judge true team performance – I think Matthew’s done some posts on that.

I was supplementing your thought rather than disagreeing with it.

Also, if you are going to look at Pythag (RS/RA)

You should at least be looking at third order wins which factor in the actual teams they played against. A lot of people (not here, but out “there”) have been using Pythagoras to suggest the M’s were “lucky” last year and not “really” even a .500 team, but by their 3rd order wins they “should have been” 83-79
…ie only “lucky” by two wins.

Of course, that was last year, which really doesn’t have any direct bearing on next year, as others have said — we have to start over with the talent on hand (both on the Ms and the teams they play against) and see how that projects from scratch, not some plus or minus delta from last year.

I love these WAR tables so much.

You guys should just post one every week, even if nothing has changed.

Bart's Evil Twin - January 9, 2010 reply

That would be cool.

A close race with the M’s having a decent shot at the division. Loving it!

It's almost as if Matthew does it at the beginning of every series.

abender20 - January 9, 2010 reply

It should be a fun season!

waldo rojas - January 9, 2010 reply

I will be very surprised if Adam Moore doesn't play in more games than Johnson.

.Taylor - January 9, 2010 reply

I love how Napoli is the Angels best player and Sociscia basically hates him

Poochie - January 9, 2010 reply

This little factoid is something that makes my heart smile.

BrettJMiller - January 9, 2010 reply

Is he their best player?

I’d be pretty sure that whatever WAR version Matthew’s using has a generic positional adjustment for catcher, but no adjustment for defensive value. In otherwords, it treats good defensive catchers the same as bad ones. If Napoli is around average as a catcher, maybe he is their best player. If not, well……..

nathaniel dawson - January 9, 2010 reply

He may not be the best but he's in the top two

They don’t really have any star players now that Figgins and Lackey are gone. Kendrick has potential, and Santana too, if he can get his fastball back.

tdot mariner fan - January 9, 2010 reply

Great job with the table

I think you guys did as good of a job remaning impartial with the Mariner’s numbers as any true fan of the team can. It is even better when considering you had to do this in the face of roster volatility from spring training position battles and promotions.

The only thing I want to note is, this race could be significantly closer than it even appears right now. Oakland’s and Texas’s rosters have considerable upside, that your projections rightly tempered in my opinion.

Oakland’s pitching could improve significantly; with Mazzaro, Duscherer, and Gonzalez having the potential (especially Gio in my opinion). Even if one of those falters, Cahill could realise some of his immense prospect potential as a GB ace, and they have Simmons and Mortensen sitting in the minors also. I realize this is partially asking everything to break right, but if 2/3 bottom starters put up a 2.5 win upgrade things get closer.

Texas has a similar situation with their offense. Teagarden and Davis both have the potential to rebound, and I think both almost certainly will if not break out altogether this year. A 2 win Davis and Teagarden would make things painfully close for the Mariners.

Crisp should be CF for the A's, replacing Buck (who will definitely not be on the roster without an OF trade);

I’d say 1.5 for Crisp. 1 WAR is too low for Barton. I’d say 1.5, but I’d say 2 before 1.

These are great posts.

Cruz and Harden look too high to me, especially Cruz since these projections are generally conservative

TheQuestforMerlin - January 9, 2010 reply

Interesting stuff

Angels fan here, like the projections and work.

Just a couple of points:

1. Brandon Wood will probably start as the 3B. Don’t know if that affects your projections.

2. Also, last year Macier Izturis was on our bench, and posted a 2.8 WAR. If you considered him the starter (at 2B) then Howie Kendrick posted a 2.0 WAR from the bench. Now I don’t know much about the Mariners, A’s and Texas’ benches but Scioscia does like to use his quite a bit. So I’m proposing a 2 WAR bench for the Angels? Considering Macier Izturis will probably start as our utility player this year.

How do you get from your WAR total projections to the win totals? Sorry if this is a dumb question, i’m not a huge fan of WAR (due its reliance on all kinds of difficult to quantify variables and on the ‘unreliable’ UZR) and consequently don’t understand the connection between WAR and final standings W-L.

You just add the WAR projections to ~48, the WAR total for a team made up of entirely replacement players.

BRKLN M'S - January 9, 2010 reply

Nitpick

48 is not a WAR total, it IS replacement level. Anything above that is WAR (on a team level).

Thank you, that was basically my first attempt to answer such a question.

I’m trying to change.

BRKLN M'S - January 9, 2010 reply

You know how 0° C is 32° F?

0 WAR is 48 wins.

Going from WAR to wins is kind of like getting your weather reports from Canada. Well, aside from the whole 9/5ths thing.

TheQuestforMerlin - January 9, 2010 reply

A team full of replacement players would be projected to win ~48 games.

The WAR is added to that to give a projection for the entire season.

Mariner John - January 9, 2010 reply

Cheers guys

So I presume this would mean the Detroit Tigers of 2003(who posted 43 wins) would actually have a squad that posted a minus WAR? That the total of their squad performed below replacement level in theory. I need to look into that.

Yes.

Or alternately they were unlucky.

Actually yes.

They were -0.6 with batting and +3 (which is very low) with pitching. Also, WAR does change year to year (which means it may have been lower in 2003, I’m not sure), but most likely they “Should” have been a ~48-50 win team, so not that far off. But yes, they were terrible. -.6 WAR with batting so impressive with its awfulness that it should get an award. It means the entire team should have paid the Tigers just to play.

And WAR for each player is independent of each other.

So there is … I don’t know if I’d be comfortable calling it luck, but “timing” involved in the difference between actual win/loss record, so one should not expect WAR to match up exactly. Only to be fairly close.

Luck will factor into that

remember that WAR is mainly a tool for evaluating an individual player’s theoretical contribution to his team’s win total. It is not a tool for retroactively calculating the number of wins a team got.

I thought it was ~46 wins.

And I can’t find anything that answers the 45, 46, 47, 48? question definitely. Anyone know an article with an exact answer?

Decatur - January 9, 2010 reply

It changes every year, which is why there is confusion.

nathaniel dawson - January 9, 2010 reply

And there are different people that have taken their shot at putting a discrete number on it and come up with different win totals.

I’ve heard every thing from 36 to 50, with the majority being mid-to-high 40’s.

UZR is fairly reliable as a one-year measure of value

the problem is that it’s not very predictive except in a much larger sample size

seattlebruin - January 9, 2010 reply

...

UZR is fairly reliable as a one-year measure of value.

I don’t think that is the case. UZR has a lot more measurement error than people think. Just ask MGL!

http://www.insidethebook.com/ee/index.php/site/comments/taking_the_uzr_out_of_uzer_error/#20

Gutierrez also looks too high

I’m not sure it’s ever reasonable project anyone at ~+15 in CF.

Guti is being regressed two whole wins.

Well, he's being projected at 2 wins less than last year, maybe...

In any case, CHONE had him at 5.4 last year, the statcorner estimate of his offense is lower than the fangraphs estimate, and it’s not particularly likely that he was actually worth 29 runs above average on defense last year even leaving aside the issue of how he projects this year.

And the guys with the highest scores should be regressed the most…

Statcorner has defense?

That's what Jeff Zimmerman's calculations at Beyond the Boxscore say (14.8 to be exact).

And Gutz had a 20.3 UZR/150 (~10.0 equivalent in CF?) in 2007 in half a season as a RF, a 26.9 UZR/150 in 2008 in 3/4 of a season in RF (~16.9 in CF?) , and a 27.1 UZR 150 season in 2009 as a full time CF. +10 in 1/2 season in 2007, +15 in 3/4 of a season in 2008, and +27 in a full 2009 in center field. +15 seems eminently reasonably for next year (although his offense should start peaking and his defense should start declining right about now, which should balance out for the next 2-3 years nicely. There’s more on this if you search for Gutz or Gutierrez and look through the past week of comments.

Decatur - January 9, 2010 reply

Here is the top average uzr/150 for the decade:

A Jones: 20
Rowand: 9
Patterson: 8
Crisp: 8
Beltran:7
http://www.insidethebook.com/ee/index.php/site/article/the_best_defenders_according_to_uzr_of_the_decade/
To me this indicates that no one, however awesome, has a true talent of +15. CHONE has him at +6 for next year.

CHONE is being extremely conservative

3 years of UZR data is around what you need to have an accurate picture of a defender’s abilities and we have that for Guti. He’s just plain better than at least 4/5 of the people you listed. +10>x>+15 isn’t a crazy projection.

Expecting him to repeat a +30 year would be, however.

OlSalty - January 9, 2010 reply

ZR had him at +7 (total, not per/150)last year. Of course UZR is better, but ZR is worth using because it uses stats instead of bis data, and that can make a big difference, e.g.:
Ichiro 2003-2007:
bUZR: +60.7
sUZR: +21
sUZR and bUZR gave a 100 run difference for the decade. CHONE is valuable because although TZ is also not as good as UZR, he uses an independent data source.

He would need to be an all time great caliber CF to warrant a +15 projection. Does 3 years of UZR data tell you that?

Yes it does tell you that because it has him around +13

He is an amazing player.

OlSalty - January 9, 2010 reply

But his point is that UZR isn't neccesarily accurate, even over a large sample size

This is due to bias in the data providers. Andruw Jones is the classic example. bUZR has him as the best fielder in the history of baseball. sUZR has him around average. Most likely, he’s somewhere in between.

Maybe...

But that’s for only CFs that have played over 500 games in the decade, and every single person on that list is now over 30, which is when they start to slow, and have been injured playing the same position. In addition, if you look at the list, there are several people at other positions that saved more than 15 runs. And there is a minimum of 500 games played, something that most young, faster players are not going to have reached. Much like it’s projections with Ichiro, Chone is known to be harsh on people that appear to be outliers, but that doesn’t mean it’s not possible, or even probable.

I think +10 is a better projection simply because assuming +15 for any player, ever, is a little too high, but if someone wants to project at +15 I don’t see any reason to think that’s not possible based on his speed and history.

I basically agree with all of that, so I'm being nitpicky to that extent.

I know .5 wins is basically nothing, but I am a little surprised that Matt Palmer is .5 wins above Vargas.

I thought Palmer was the definition of replacement level (0 WAR) and while Vargas is nothing special I thought he was closer to an RR-S type, average but with some decent upside.

MFAN - January 9, 2010 reply

Two things:

1) How does Johnson project 0.5 WAR more this time?
2) Why wouldn’t Kotchman project to contribute more at 1B than Hannahan?

uw_chris - January 9, 2010 reply

It's pretty much impossible for a catcher to be replacement level

What?

Seriously, is this an inside joke I missed somewhere or are you serious?

Catchers get a tremendous positional adjustment.

And they do not currently have a way to give /- based on defense, so the only way for them to be replacement level is with a bat that is so bad as to be almost hilarious. Dioner Navarro was replacement level last year, and he had to run out a wOBA of .258 (58 wRC) just to do it. An, of course, when someone is that bad, they are usually replaced with someone better. To put that in perspective, Rob Johnson had a -12.1 with the bat (67 wRC+) and was still worth almost a half a win.

We can assume sucky defense, but we can’t really measure it, so it’s almost always excluded, and without a defensive contribution and with a positional adjustment that is so damn high, you have to be all kinds of terrible with the bat to be replacement level – so bad not even someone that loves “catcher defense” can justify keeping you on the roster. Pretty much a big black hole.

The statement "it's pretty much impossible for a catcher to be replacement level" is equivalent to saying that replacement level catchers do not exist in plentiful quantities.

This seems to disagree with the whole concept of replacement level.

No it isn't. It's equivelent to,"it's nearly impossible for a catcher to be bad enough to be considered replacement level"

There’s no need to mince words.

So then we all agree that replacement level catchers DO exist in plentiful quantities?

How is it near impossible for a catcher to be replacement level then?

By "replacement level"

I am referring to -2 wins below average version. The problem with that is a true replacement level catcher is probably better than 0 WAR, because if you look hard, there are really no 0 WAR catchers out there.

A "true replacement level catcher" would provide 0 WAR, just like at every other position.

You’re contradicting yourself.

Assuming FanGraphs positional adjustments are correct

By definition, a "true replacement level catcher" provides 0 WAR.

WAR = Wins Above Replacement.

0 FanGraphs WAR =! 0 WAR

For that to be true, the positional adjustments would all have to be perfect, and we know that they aren’t.

I think we would all agree on that.

And then the question becomes are we (or in this case, Matthew) making projections as to how players will end the year in fangraphs WAR or what we’ll call “True WAR”? I would assume the latter.

And yes, by the way, it is equivalent to saying that.

By saying that it’s nearly impossible for a catcher to be replacement level you’re implying that few catchers are replacement level, which is to say that they are rare and do not exist in plentiful quantities.

We can measure the value of passed balls/wild pitches and cs% quite easily.

Which are two substantial parts of catcher defense. We cannot measure the value of pitch framing (yet), ability to receive a throw and block the plate to runners (likely amounts to maybe a run or two difference between the best and the worst over the course of a season due to the infrequency of the event), or call a game (likely never). Am I missing anything?

Passed balls and CS are HEAVILY affected by the pitcher

Really the only way to adjust for that is via a WOWY method, and none of the systems currently do that.

Catchers can be replacement level in skill.

It is much harder for catchers to be replacement level within WAR calculations.

Since 2002, there have been an average of 6.5 catchers/year with WARs between -.4 and .3

In 2009, of catchers with 200+ PA, 6 had WAR’s of .3 or lower. In 2008 it was 7 players. 2007 was a good year with only 3 , but 2006 saw 7 as well. 2005 had 4, 2004 saw 9, 2003 had 6, 2002 hit double digits with 10. There are an average of 6

Whoops, ignore that last fragment.

Well, first of all you should make your cut off higher.

200 PA is pretty much a bench catcher in a slump. If you raise it to 400 PA (which assumes a starting catcher), you’re see that there is basically 1 per year at most that is replacement (0 WAR) or lower. 1 per year is very, very few – and if you look at HOW bad they were, you’ll see that they were a level of bad that even Rob Johnson probably won’t manage, especially at his young age. 0.3 is not replacement level. 0.0 is replacement level. And if you look at how bad the catcher had to be to get 0.3, you’ll see how unlikely it is to assume that Rob Johnson is going to be that level of suck. Bengie Molina was exactly 0.0 WAR in 2002, and he had to hit an amazingly abysmal .596 OPS. I mean… it’s hard to be that bad. And though it seems like there is a catcher doing that every year, you almost never project any major league player to be sub-.600 OPS.

Rob Johnson has a career OPS of .589.

I wouldn’t project him to hit that badly going forward, but just saying…

Also, I’m assuming fangraphs WAR has a decent margin of error. Fangraphs WAR likely has a decent margin of error from “True WAR”. I have no problem labeling a player who reaches a measly .3 WAR as likely replacement level.

.

Rob Johnson has a career OPS of .589.

Okay, no. I know you don’t believe this. His only season beyond like 10 games was last year, and he had a .615 OPS (if we want to use OPS), and he was STILL worth ~.5 WAR and he only had 290 PAs. Over a full season, that’s 1 WAR, and that’s with his awful .274 wOBA.

Okay, I'm going to cut this conversation off early.

“Why Does Rob Johnson have a WAR of 1”

Because of the tremendous traditional adjustment of catchers, Rob Johnson has a to hit a measly ~.274 wOBA over a full season to reach 1 WAR.

“Why is He Only Projected to Get .3 WAR on Fangraphs?”

Chone has him at 379 PAs, or about 1/2 a season. Chone also has him at -5 fielding, which I don’t understand, but okay. I GUESS we can give him -5 fielding, because it’s probably accurate, but meh, we don’t usually quantify catcher defense in these projections, so he should be at .8 WAR for just half a season. If we assume full season, even with as terrible as he is, he’ll be at .6 WAR if you think he’s -10 at defense, and since he’s probably not THAT bad, 1 WAR is a safe estimate.

What if He Only Plays Half the Season and is Replaced by Adam Moore?

Good question! But the answer is: Doesn’t really matter much. Adam Moore has probably the most retarded Chone projection I’ve seen – They give him -14 runs on defense, as a catcher, in half a season. Yeah, no. He’s not that bad. But he’s still not going to be great, so his projection will probably be ~Rob Johnsons, and they will both equal roughly 1 WAR as a full time player, and either way he gets 1 WAR. No matter how you slice it, Rob Johnson at a full season or a complete split between Rob Johnson and Adam Moore – either way, it’s probably about 1 WAR for our full time catcher.

Wasn't calling you out here Term X - just didn't have any real place to put this so I put it here.

No problem, that needed to be posted somewhere, great summary, one that I find little to disagree with.

I was just taken aback by the “It’s pretty much impossible for a catcher to be replacement level” and couldn’t help but ask for clarification.

I think the real question is not whether one can be a 0.0 replacement level catcher.

But whether or not replacement level should be altered since we can’t quantify defense. It is very hard for a catcher to be 0.0 WAR by most current calculations, simply because of the huge positional adjustments. But that implies that the replacements are better than replacement level and that can’t be true logically. Replacement level should be the level that replacements are.

Once they can calculate catcher defense though, I think this problem will be solved.

I think we are talking about different things.

You are talking about how there are catchers that suck enough to be replacement level. And that’s true. But holy crap do they have to suck – an unexpected level of suck. I’m talking about how for projection purposes, the chances of projecting a 0 WAR catcher are virtually nothing, because for that to be possible you would have to be expecting the catcher to be one of the top 3 to 5 worst hitter in baseball. There were only two players last year that hit bad enough with 400 PA that they could have been 0.0 WAR catchers, and it just so happens that one was actually a catcher.

Look man, all vivaelbirdos was saying is that

with replacement level and positional adjustment a catcher has to be a whopping 37-40 runs below average with the bat to be replacement level. That’s unbelievably shitty, and precisely because there are so many catchers hanging around every organization when a guy sucks that much you DL him and plug in another.

Look at the list of backup catchers and the value they provide. The majority of backups have positive WARs on the year, and that’s with very limited playing time. Johnson is a virtual lock to produce ~10 runs if he shares his duties 2:1 with Moore, but I think Seattle’s catchers will total something like 2 wins, just as they did last year, terrible as they were.

Time out.

I never disagreed (or even said anything) about Rob Johnson’s projections. I was just trying to figure out what exactly he meant by “it’s just about impossible for a catcher to be replacement level”.

Though if you want to bring Johnson into it,

Saying “it’s nearly impossible for a catcher to be replacement level” doesn’t answer the question “why did Rob Johnson jump from a projection of .5 WAR to 1 WAR”.

I can't answer for Matthew

many of his projections changed a little one way or the other since last time.

And in this case I think the original answer, that a catcher has to be epically bad to be worth so little over a full season, addresses the change. Johnson is not a good hitter but so long as he can keep his wOBA around .300 he’ll be worth >1 win over a catcher season.

I think what he meant was that you basically have to be an average AA hitter to be a replacement level catcher

Damn those fonts are snazzy.

http://www.baseballprojection.com/2010/SEA2010p.htm

Lee is projected too high

Lee at 5.5? His weighted average WAR on FanGraphs is 5.2. Plus, over the past 3 years he’s been running a ~7 HR/FB that’s bound to go up. That’s before regression and aging, and not even considering injury risk. CHONE projects his at around 4.5 WAR.

Dude.

Again, these are just my projections. This is a gross estimation.

Come on, man. It’s in the article.

So because he says they are gross estimations, we shouldn't be allowed to criticize them?

I think you're taking it the wrong way

Criticisms are welcome as long as they’re respectful and logical. I think he means don’t get too worked up over .5-1 WAR since this is still a rough estimation. I’m sure Matthew will breakdown why he gave Lee a 5.5 WAR once he has the chance.

Scrupio - January 9, 2010 reply

Well that's fine

In the meantime, I don’t think it’s inappropriate to say that I think his estimate of Lee is off.

nathaniel dawson - January 9, 2010 reply

He seemed pretty reasonable to me.

Just giving his considered opinion in what appeared to me to be a constructive way. And it’s not as if this guy is some yokel.

No, it means they are admittedly 'gross' estimations,

and so nitpicking should be minimized. Criticizing should be made in broad strokes, or not at all, like the mention of Mathis vs. Napoli’s playing time. One is obviously the superior player, but gets PT taken due to irrational manager hatred.

Criticism is great, especially in M’s blogland.

misterjonez - January 9, 2010 reply

No, it's okay.

This was my mistake. I read your criticism as “You’re wrong, his projections are the following:…” rather than "I think you should change his projections thusly…’ I would personally prefer that you start your sentences with “I think” instead of “You’re projections are too high” but that’s a minor preference.

To respond to your general thoughts:

1) If he has been running a 7 HR/FB for 3 straight years and is moving to Safeco, I think it is safe to assume that there may be a skill involved in keeping those home runs in the ballpark. You may disagree, and that’s fine, but usually regression occurs the following year, if not the year after. 3 straight years would suggest he is doing something differently, and even if he is not, Safeco is going to be in his favor.

2) Chone’s projections are based on a 3 year weighting system, and it is pretty clear he figured SOMETHING out these past two years, considering the huge difference as well as the injury. The pitcher has been worth about 7 wins for 2 years straight. 5.5 is a fair estimation. In addition, Chone is going to regress his HR/FB considerably – more so than they probably should. They are also going to regress is GB%, which again appears to be fairly constant. Even if you regress them a bit, 5.5 is still a pretty decent bet. You shouldn’t regress them as much as Chone does, especially because they are also weighting 2007 which is going to lower the estimations to a probably unfair starting point.

3) The difference between 5.5 and 5.2 is fairly minimal, and I think because it is safe to assume that Cliff Lee doesn’t need to be regressed as hard, and that his 3 year weighted WAR is probably too low, raising it by just .3 doesn’t seem like a stretch.

You're right that I should have added an "IMO"

As to your points…

1) In 07, his HR/FB was around league average. In 08 and 09 it was about 5%. It’s possible that he actually developed a skill to post lower than average numbers in that category, but A) you can’t tell from those two data points (HR/FB numbers take about a half a career to tell you anything meaningful about the pitcher), and B) it would still likely a be a lot higher than the 5-6% it’s been over the past 3 years.

2) I’m not really sure exactly how CHONE projections are based. It probably uses at least 3 year’s of data, but I’m betting it regresses and weights each component individually. So things like GB, K and BB rates will have much less regression and a heavier weighting of more recent data then stuff like BABIP and HR/FB rate.

3) The FanGraphs weighted average doesn’t include regression to the mean or aging or injury risk – so it’s obviously going a bit inflated.

At any rate, I think it’s a reasonable debate as to how good Cliff Lee is going forward. I think a lot of signs point towards a more conservative projection but I could see the case for the other side.

Yikes, you mean we traded for a guy who gives up 7 home runs on each flyball?

Not even Brandon Webb could get away with that.

Okay, General question.

WAR is supposed to be somewhat context neutral, correct?

If so – Does how suited your team is to its park affect estimated win totals? We’re getting more and more lefties (both pitchers and hitters) – so even if our WAR is, say, 40, wouldn’t it in some ways be more because we’d play 81 games in a park more suited for our skill set than a generic team?

If not – Do the projections reflect estimated player contributions in the new ballpark? If Cliff Lee was 5.5 in a hitters park, should he be 6.0 now that he’s in Safeco? Similarly, though Safeco isn’t actually that great for lefties either, couldn’t Kotchman get a park boost?

Doesn't have to be our team either.

Or our park, but more in the generic scheme of things – if you plan your club to hit really well in your ballpark, will their projected wins be greater than their projected WAR?

Fangraphs' WAR for pitchers has a park adjustment

At least that’s my understanding based on Dave’s explanation at Fangraphs. But it’s a pretty simplified park factor (it’s not different for RH and LH, for example) and he notes there’s a fair bit of noise so they’re using a five year regressed value.

They’re using the same value (I believe) for the adjustment they’re using to calculate wRC+ for hitters, but that doesn’t factor into WAR for position players.

But it’s Saturday night and I’ve been drinking so I won’t even try to think my way through answering your larger question.

Two things...

1. I want to see a projected batting order. Anybody, anybody, bueller?

2. Were we even thinking about getting a third really great starting pitcher? Was anybody out there? IS anybody out there? Are the three guys we have penciled in there now… well what do we think of them? I know they call all pitch but can they excel?

Emperor_Doom - January 9, 2010 reply

We might be

Probably not “really great” because we don’t have a whole lot of money left, but a pitcher who’s cheap with upside.

OlSalty - January 9, 2010 reply

Jack Z

has certainly shown that “out there” is kind of a relative term. Chances are, if we get someone, it will be someone no one even thought of. The obvious candidates, though, are Sheets, Harang, Bedard. First two may be out of our price range, third may be opposed to coming back.

Fuckmikereilly - January 9, 2010 reply

Batting order

Wakamatsu seems to be a fan of mixing things up, and given a roster with so much positional flexibility I’d expect to see all kinds of shit in the lineup cards.

Yeah beyond 1. Ichiro and 2. Figgins it's anyone's guess

Scrupio - January 9, 2010 reply

Oooo...guessing!

Maybe this will suffice Emperor_Doom…

1. Ichrio
2. Figgins
3. Bradley
4. Lopez
5. Gutierrez
6. Kotchman
7. Saunders
8. Johnson
9. Wilson

With the possibility of 3 & 4, 6 & 7, and 8 & 9 switching spots with each other.

SethGrandpa - January 9, 2010 reply

Thanks

For the Record I want to sign Bedard again. let the guy have his chance. He could be a #1 starter in the 3 spot. I know he’s had his problems with injuries but that can’t last forever… well it can, but it would be worth the stretch to my mind. The reason?

If it doesn’t work out it doesn’t doom your season.

if it does work out. Hello division title and playoffs.

Emperor_Doom - January 10, 2010 reply

I guess Kotchman fifth

and whoever of Lopez/Guti is hitting better fourth. Otherwise this looks about right, but of course depends on how the season goes for everyone.

I think we put Gutierrez at three, Bradley four then Lopez five, but I would keep the rest the same.

InSpokane - January 10, 2010 reply

I noticed that in your last calculations You put Hannahan down as a +1 1B and Carp down as +1 at DH.

Here, you put Kotchman down for +1 at 1b. I have to imagine he’s better than that or the other two projections were overestimates?

greymstreet - January 10, 2010 reply

I actually feel pretty bullish on Kotchman...

Sure he’s never going to be a “proto-typical” offensive player at first, but I wouldn’t be at all surprised if he went out and had a .280/.350/.850 type season.

And if he doesn’t? You’re down Bill Hall and a C-’spect, and you move onto someone more productive (Carlos Pena?) in July.

I can’t say I have complete faith in Zduriencik, or some ridiculous notion that he can do no wrong, but I do feel he earned the benefit of the doubt with all the “suspect” signings from last year, and I have faith that he did his homework when he decided that Casey Kotchman was deserving of a starting job.

Adam B - January 10, 2010 reply

That 850 should be an 800.

I’m not THAT bullish.

Adam B - January 10, 2010 reply

Zduriencik knows Kotchman from back in his scouting days

When Kotchman was in highschool in his scouting region, apparently. So he may have some reason to believe that Casey will outperform his projections (and live up to his better past seasons, if not the potential that led him to be picked in the first round) given regular playing time in a single organization (and good health).

wandergeist - January 10, 2010 reply

So CHONE projects Adam Moore to be -14 with his glove. While only playing half a season.

This must be payback for signing its namesake.

That just cannot be right

although the strongest support I can muster for my statement is the org is very high on him all around ergo he can’t be that bad.

WooHoo!

Pop the corks baby! 2010! Can I get playoff tickets yet? What’s the magic number?

Rich Langford - January 10, 2010 reply

162

seattlebruin - January 10, 2010 reply

I think it is 143.

Oh hey! That's not right.

nathaniel dawson - January 11, 2010 reply

I believe 42 would be the correct answer

M's have 4 of the 5 players rated at 4+ WAR in AL West

That just struck me as interesting. Ichiro, Guti, Felix, and Lee are all rated as 4+ WAR. The only other AL West player at 4+ WAR is Kinsler at 4. Assuming that your estimates are accurate, I wonder if that means anything. Perhaps the M’s subscribe to more of a Stars and Scrubs style of roster construction than do the other AL West teams? Or it might just be somewhat coincidental, I dunno.

urchman - January 10, 2010 reply

Texas has a lot of young players who are coming into their prime

they will have several 4 win players in a few years. The Angels have had a number of 4-5 win players recently who have left the team or hit decline. A four win player is not quite a star in the stars and scrubs sense. I think if the M’s had three 6-8 win players and a bunch of role player types (hellloooo future Cardinals teams) then you could call it that. I think it’s just a sign of a poor division.

I'd hardly call the AL West a poor division.

Only one team projects to be below-average, after all.

katal - January 10, 2010 reply

The advantage to the M's in this case is that it's much easier for them to upgrade than for the Angels.

The Angels have alot of average to slightly above average players – only the fifth start projects as below average. To add 2-3 wins they have to add a borderline all-star player or add a couple 3 WAR players. The Mariners can add the same amount of wins by just sprinkling in a few league average players (C, 1B, LF, SPx2). Texas is somewhere in between.

Terminator X - January 10, 2010 reply

To answer a few things:

Johnson changed because I underestimated how hard it is not to get 1 WAR out of catching.

I wasn’t really thinking Hannahan would be a 1 WAR first basemen. Think of it more like a listing of the current starter and projected total WAR, for the position, not the player. That is, I figured the Mariners with Hannahan and Carp and whatever would get about 1 WAR out of their first basemen. Kotchman doesn’t really change that much.

On Guti, first of all, I think +6 is too low on defense. Secondly, and more importantly, CHONE projects him for 511 PAs. He got 630 last year and missed 10 games. CHONE is giving him low playing time because of his 2007/8 playing time totals, which are no longer relevant. They weren’t injury-related. Even assuming the roughly standard 85% playing time for a projection, Guti is missing about 40 PAs, which is another 3 or so runs.

Matthew - January 10, 2010 reply

Your Texas projection is pretty much in the shitter now that Khalil Greene is a Ranger