Seattle Sports Insider

What am I doing wrong?

Posted by

1/21/10

Let's make some loose estimates for how many wins are on this roster as it stands right now. If we add significant players, we'll update these projections. I'm going to be very conservative with the older players and with unknown quantities (the kids and guys who haven't gotten much of a track record yet). Think of this is a roughly lower-middle-range projection:

CATCHERS: +1.1

Rob Johnson (220 PA) - +0.3
Adam Moore (270 PA) - +0.6
Josh Bard (140 PA) - +0.2

FIRST BASEMEN: +1.2

Casey Kotchman (480 PA) - +1.0
Mike Carp (1B/DH) (250 PA) - +0.4
Tommy Everidge (140 PA) - -0.2

SECOND BASEMEN: +3.0

Jose Lopez (640 PA) - +2.8
Matt Tuiasosopo (2B/LF/3B) (60 PA) - +0.2

THIRD BASEMEN: +4.7

Chone Figgins (640 PA) - +4.2
Jack Hannahan (3B/SS/2B) (140 PA) - +0.5

SHORTSTOPS: +1.4

Jack Wilson (500 PA) - +1.7
Josh Wilson (75 PA) - -0.3

LEFT FIELDERS: +4.1

Parker Brothers (LF/DH) (500 PA) - +3.2
Ryan Langerhans (OF) (220 PA) - +0.9

CENTER FIELDERS: +4.5

Franklin Gutierrez (650 PA) - +4.5

RIGHT FIELDERS: +5.0

Ichiro Suzuki (720 PA) - +5.0

DESIGNATED HITTERS: +0.0

Ken Griffey Jr. (300 PA) - +0.4
Extra Scrubs / Pitchers (200 PA) - -0.4

TOTAL OFFENSE/FIELDING (5945 PA): +25.0

STARTING PITCHERS (1000 IP): +16.5

Felix Hernandez (240 IP) - +6.5
Cliff Lee (200 IP) - +5.2
Ryan Rowland-Smith (160 IP) - +2.3
Ian Snell (175 IP) - +1.5
Doug Fister (120 IP) - +1.1
Jason Vargas (65 IP) - +0.2
Luke French (40 IP) - -0.3

RELIEVERS (450 IP): +4.5

David Aardsma (65 IP) - +0.9
Brandon League (90 IP) - +1.8
Mark Lowe (75 IP) - +1.0
Shawn Kelley (70 IP) - +0.5
Sean White (55 IP) - +0.3
Nick Hill (45 IP) - +0.3
Random Scrubs (50 IP) - -0.3

PITCHING TOTAL: +21.0

TEAM TOTAL: +46.0

LAST YEAR: +37.0

PROJECTED WINS: ~94.0 (?)

OK...so where did I go wrong...because that seems stridently optimistic. Taro, anyone? Which of these projections seem unreasonable to you.

This seems to occur every year when I attempt to add together the pieces of a team and make a projection. What is the error here? I know the Mariners aren't a 98 win team on paper...so what gives? I projected an UP season for only Jose Lopez and Brandon League (relative to career arcs) and some minor upticks for guys like Parker Brothers and Shawn Kelley relative to the most recent seasons. All of the kids I have as being barely above replacement level. I cannot find an overly aggressive projection here.

Comments

One possibility...

just at a glance, so check me ... maybe part of it is negative contributions by players who are only on the roster a short time?
You've got -0.3 for random scrub pitchers and -0.4 for random DH's/etc...
However, in 2009, the Mariners got close to -30 runs (vs RLP) just from Betancourt, Cedeno, Hall, Burke, Woodward and Shelton.
They also got -10 runs (vs RLP) from Garrett Olson and another -10 from Corcoran, Stark, and French.
That would account for about 5 wins that the transition feebs cost. Don't know whether that addresses the question.
...............
Still and all, the M's have some pretty robust run prevention. If their offense actually gets 7 wins from LF and 3B, they stand to win 90+.
Great post Matty :- )

And if so...

... it could easily be that in 2010, the M's won't have nearly as many as these traveling banjo players stopping in Safeco. Could well be that the M's have figured out their 25 to a much greater extent.
Which could lead to an invisible 3 wins or so vis-a-vis 2009.

The Ms have pretty much

The Ms have pretty much flushed their sub RL players and have a bunch of RL scrubs ready and available in AAA. Its likely one of the hidden values this season.
I just did my own projection over at MC. I was conservative, but still came out at 90+ Ws. What? :-)

Scary good

It isn't clear that you'd get the same results by Pythag, though you'd think you should. Defense should be improved somewhat over 2009, and a healthy Lee >> half season Bedard. We have also filled several black holes in the lineup. I think we are still missing 200 solid innings of a #3 SP that we had with Wash in 2009. Take care of that and look out.

And I strongly disagree with taro in two places.

1 Jose Lopez...his projections had Lopez a full win worse than mine...I think he's slowly improving defensively and I think he's just as likely to have a career year offensively as hit at all worse than he did in 2009...so my mid-projection suggests small uptick...him being the ONLY player my projections suggest such an uptick from.
2. His projection for Kotchman is incongruously worse than Kotchman's career performance to date. I don't like Kotch either, but I looked right at his current WAR lines and projected a mild downtick defensively and constant offense.
The role players and RLPs we'd need to use if injuries occurred this year are also all likely to be better than the RLPs we used in 2009.
I might be too bullish on Parker Brothers...not sure. Depends on how serious the problems in 2009 were or whether there was a psychological element that can be addressed.
A big part of the reason I didn't see a lot of negative performances is because I couldn't find playing time for the guys who would be negative performers. For example, I tried to be very pessimistic about how many innings we could get from RRS/Snell and I still got >700 IP from our top 4 starters rather easily, leaving me very little room for RLP starting pitching. Last year, the Mariners have one guy who threw 240 innings and then no one else with half of that total. RRS is going to start the year healthy this time. We have Ian Snell healthy and ready to pitch this time. And of course you can count on Lee for as many innings as Bedard and Washburn pitched COMBINED...so how are we going to need more than 250-300 innings out of RLPs this year? And do we really expect guys like Fister to be as bad as the RLPs we used last year?
The bullpen has the same story...we've got four guys who could easily log 300 innings assuming they are all healthy. Aardsma, Kelley, Lowe and League...all capable of 75 or more quality innings if need be, and I was pretty bearish on Aardsma in my projection. That leaves 150-200 innings for a bunch of guys to fight over, at least a couple of which are likely to be better than RLP and get the majority of those innings...so...I'm having a hard time seeing a lot of downside from my own bulllpen predictions...the only projeciton you might look at as optimistic is Sean White getting 55 innings (he might be more seriously injured than we've heard so far and throw most of those innings over to RLPs, which would hurt).
I had to work pretty hard to find bench PAs too...My starting nine accounts for 4700 PA (last year's Mariners got 4101 PA from the guys who logged the most innings in the field at each of the 8 positions and Griffey at DH - this year's projections send more PA to the starters in LF and SS and that is the only real difference...both of which make sense...SS is now a starter, not a sub RLP pile of crap, and LF is also now a starter with Bradley slated to take over essentially full time rather than a mix of Endy Chavez and Wlad Balentien and Michael Saunders and Ryan Langerhans)...that leaves me only about 1500 PA to hand out to benchies...I gave 590 to back-up 1B/DHs, 360 to back-up catchers, 220 to Ryan Langerhans and 275 to pack-up infielders (Josh Wilson, Jack Hannahan, Matt Tui)...those are the four positions we have on our bench and I assumed everything but Langerhans as 4th OF was very fluid and subject to change.
I could project significantly worse results for the scrub pitchers and scrub hitters and still find 93 or 95 wins on this team because there isn't going to be a lot of PT for scrubs unless we start getting hammered with injuries.

Slight Error, I believe

Matt,
It looks like the starting pitching adds up to 17.0, not 19.0, making the total pitching number 21.5, not 23.5.
God bless, Michael

Ooh...thanks! Missed that mistake.

Hey Michael...thanks for catching that one...I counted RRS twice when I was doing the mental math (d'oh!)...I fixed it in the post.
Still 96 wins seems a little too optimistic to me. Taro's 90 win projection makes more sense...but the only way to get there is to be EXTREMELY negative about all of the scrubs.

In general, I think you're

In general, I think you're pretty consistently taking a 75th percentile projection rather than a median one. Every year some players will over-perform and some will under-perform, which is why people tend to be more conservative. It's not that any one projection is unrealistic, it's just that they're all a little bit optimistic. Teams rarely have everything go that right.

I don't know how that can honestly be true.

I projected exactly four eplayers to do better in 2010 than they did (in terms of production rate) in 2009. Milton Bradley, Jose Lopez, Shawn Kelley and Brandon League. In each case, the up projection was based on a critical observation (Lopez: 4 consistently great offensive months in a row to end the year...showing his renewed focus on driving the ball when he gets a cripple pitch, Parker Brothers: coming off a down year and entering a better clubhouse, Brandon League: major second half performance spike as he learns a new devastating off-speed pitch, Shawn Kelley: Lost a chunk of time to injury and wasn't quite the same pitcher for a few months after his return)
The rest of the guys on the club I projected to either hit the career averages (which is BY DEFINITION a medium projection) or do worse (if they were aging or I felt there was reason to question what they'd do going forward).

I penalized Kotchman for his

I penalized Kotchman for his consistently awful baserunning + GIDPs. Once you go there 1 WAR is about right unless he has a good offensive year.
Lopez... Its been beaten to death so I won't go there. I think he goes back to his usual 2 WAR this year.
Otherwise I kind of agree with you that I might have been slightly pessemistic on the scrub players.. Its interesting because Z has done such a good job of plugging in RL players in AAA that we may make up 3 Ws or so from last year on that alone.
I think I may have also been a tad optimistic on the "health" of our top 5 players. If they can avoid the 60-day DL, they'll produce around those figures or even better. Any long term DL stints to 2 of those players could lead to a loss of anywhere from 2 to 3 Ws.
This team reminds me of the '03 Mariners that pythaged 99 Ws due to rediculous D. The '03 team had a better offense, but less high end pitching.
Intutively I thought the Ms were at around 88 Wins, but going over the data 90 Ws seems more logical (barring a disaster injury to one of the 5 stars) and Z likely isn't finished yet.

I'm witchoo on this one Matt

I think the 96 wins that you've got, minus 3-4 wins by short-term players putting up terrible lines, leaves you around 90-92 and that's fair.
This club has added significant wins from last year, notably at LF, 3B and SP2.

OK...that's actually a good point.

Kotchman's baserunning does cost the team runs that don't show up in his personal WAR...so that's a good adjustment. I might be too optimistic on Carp but he hit so well in short sample in '09 that I'm not inclined to think he's an RLP this season. The back-up SS I might be too optimistic about though. I can't find places to cut wins from the pitching staff though...that's 21.5 WAR hard on the barrelhead barring any injuries to King Felix or Cliff Lee.
And yes..this team will be like the '03 club in terms of historically great team defense...but with better front end starting and worse back end starting (and a better bullpen).

You forgot SS.

Last year, the Mariners got like -1.5 WAR from shortstop. This year...+1.5 is the minimum I'd expect. That's a 3-win swing.
Last year, the Mariners got about 1.5 wins from LF. This year...it's more like 3.
Last year, they got about 1 win from third base...this year...4 wins easily if Figgins stays healthy.
Last year, they got 16.2 pitching wins total...I'm having a hard time coming up with a number lower than 20 no matter how negative you want to be about the scrubs.
The club is at least 10 wins better than last year on paper and last year's club was a .500 ballclub (on paper).

I made a few tweaks to the projections.

I backed off on my projections for Kotchman, Carp, Josh Wilson, Jason Vargas and Luke French. Yielding 2 more wins shaved off the projections...I don't think there's a lot of room to come down any further unless you start assuming injuries. I really think this is a 91-94 win club (depending on how much we need to use sub RLP scrubs) and it goes lower than that only if we start having major injuries to key players.

Not stars - but scrubs

GREAT thread.
Like you, Matt - my reaction to the 98 projection was a knee-jerk, "Matt's living in fantasy land, again," moment. But, there really isn't any one glaring obvious flaw in your assessments, (especially after your rework). But, I think the real GEM of insight in this is Doc's guess that you might be over-projecting the scrubs.
There is a VERY legitimate point that the 2009 Mariners were (roughly) 2 games under at short - and another 2 games under with the back-end pitchers, (Olson, Corcoran, et al).
Remove the NEGATIVE WAR performances from 2009, and the club adds 4 (or more) wins to their '09 total. Of course, you still have a slight disconnect in that the club won more games than it 'should have' (based on run differential or linear weights -- though the total number might be different). And, WAR is spreading out credit for those extra wins amongst everyone, (even the 2009 dregs).
But, I think if there is a stock problem with projecting systems - it is that they focus really hard on the stars - the regulars -- and typically ASSUME RL performance from everyone else. The problem here is not that the club lacks RL fodder in AAA. The problem is TIMING. You don't replace the sub-RL players on ANY team until *AFTER* they have performed under RL. You cannot "project" the EOF implosion -- or the Richie Sexson implosion. But, they happen.
And while Z has been MUCH quicker to move to fix situations, it can be even more detrimental to panic at the first 3 week slump. Cedeno was the fallback for Yuni - and likely viewed as a "relatively safe" zero WAR replacement for Yuni -- his glove likely making up for his projected offensive deficiencies. Obviously, that wasn't what happened.
I have stated previously that the Bavasi teams were completely devoid of any bench help at all, and this was a major problem. BENCH hitters combine for ~ 700-1400 PAs per season. That's one or two full players worth of PAs. And, it's almost "anti-leveraged". The bulk of those ABs are often forced by circumstance, (mostly injury). The 2007 surprise team only were forced to spend about 700 PAs on the non-starters. They were unusually lucky in terms of not actually HAVING to play their RL players.
The focus of fans and analysts has been (and will continue to be) on the 700 PAs you get from your Vlads and Ichiros. But the 700 PAs you get from you Cedeno's and Hannahan's and Josh Wilson's matter just as much. And in many, many, many cases, it is the 'invisible' performances from these combined offensive slots that tip the scales.
Why did Anaheim beat their pythag and linear weights results so badly in 2009? I'm thinking it might be related to only spending about 500 PAs on non-regulars. They had 10 players with 100 OPS+ scores. They ended up with about 250 total PAs from players below 80. The problem with Seattle's 2009 club was NOT that they needed another Branyan in the lineup -- it was that they needed a lot FEWER 60 OPS+ PAs. It was the COMBO of Johnson, Yuni, Wlad, Cedeno, Hall, Saunders, Jack Wilson that destroyed the offense.
So, the question for 2010 is NOT "how will the stars perform?" The question is "how will the SCRUBS perform?" Personally, I'm getting more optimistic, because Z has been steadily working to remove the worst players and replace them with cheap-average guys. That has ALWAYS been the fastest and most efficient way to improve a baseball team.
Oddly, is 2010 goes really well, The Cliff Lee deal will likely be lauded as the 'key move'. In point of fact, it is precisely moves like Langerhans and Hannahan and Kotchman that allow the club to improve w/o risking the major gaffe that cripples the club for the next 4 years.

So the question is...

What does the average big league club get from its 1400 least "regular" plate appearances?
IOW, if you take the top 10 PA getters on each club and set them aside, what is the average WAR for the rest of them?
Right now, I have the top ten PA getters accounting for 4950 PA, leaving about 1050-1250 PA for the remainder of the scrubs. I have those scrubs combining for 1.2 WAR in 1100 PA. Is that too optimistic? Does that reflect a roughly average major league bench? I might be 1.2 wins too optimistic if the assumption in WAR is that bench players are worth (on average) 0 WAR. The scrubs on this club are, IMHO, at least MLB average at this point, but we won't know that for sure until we see them play. As you'll note, though, I am not overly optimistic about how much time the bench will get. I've made that mistake in the past and projected the regulars to get all but 800 PA or something like that...this time, I have them getting 1100 PA which is roughly average and accounts for the high durability of Ichiro, Gutierrez and Lopez and at the same time, the low durability of Wilson, Bradley and Kotchman (and of course, the catchers).