While I understand the argument for WARP dollars -- I do NOT agree with the over-simplified math to get the per-win figures.
Back in 2003, when the Tigers won 43 games -- they had from almost any statistical perspective an entire TEAM of replacement level players - and they produced replacement level wins. That team STILL had a salary of something like $43 million.
As has been shown by the Tigers ('03) and the Mets ('62), a team full of AAA players will STILL win 40 games.
The most wins ever (116), means that (in theory), the best team in the history of baseball STILL lost 46 games.
One flaw in the $ per win approach is that it treats the calculation as linear. It ISN'T. It cannot be -- otherwise the $16 million payroll 1999 Marlins should've won 8 or 9 games instead of 64. The Yankees, on the other hand, added more and more payroll, (AND production), yet maxed out on wins. Why couldn't they just SPEND their way to a 150-16 season? Because baseball doesn't work that way.
So, fine -- let's factor out the unwinnable and unlosable games -- so we've only got the +21/-21 games in the juicy center that we typically see. Well, if you're diving your $90 million payrolls by only say, 45 games, that's only $2 million per win/loss. If you factor out the $40 million of base salary (our Detroit example), then it's closer to $1 million.
Then again, if you use true league min, (25 guys at $400k), you're looking at a minimum salary of about $10 million.
My point is not to say that X method or Y method is superior -- but to point out the reality that you can create about 100 different "reasonable" methodologies for assigning dollar value per win with vastly different results. And IMO, most of this work ends up obscuring rather than clarifying anything.
What is the league AVERAGE salary? I don't know. Do you?
I know that because of the way the salary rules work, during years 1-6 of major league play, salaries are drastically suppressed. One problem I have with "replacement level" math is that "replacement level" may be very useful to a statistician, but it is meaningless to the typical fan. AVERAGE -- that he can grasp. He can look at the league average totals for all kinds of things and immediately understand whether someone is above or below average -- (though finding that data on a per position basis is much harder).
And, of course, the salary escalation system automatically skews what the value for a "replacement level player" actually is. The Tigers had a plethora of replacement level players making MUCH more than league minimum.
Anyone who understands the basic economy of baseball knows that it is completely irrational to attempt to compare players in years 1-6 to players in years 7+ based on DOLLARS. It's Euroes and Yen.
In one dollars for dingbats writeup I saw recently, they noted a figure of $14 million being "average". It was unclear as to average what. My own calculation (25 players timex $14 million = 350 million average payrolls. Wait - that couldn't be correct).
Well, maybe if they only counted the 9 starters, 5 starting pitchers and a closer (15 * 14 = 210 million -- hmmm, that can't be right either).
The problem is not that the system is wrong -- or poorly constructed. The problem is in pretending the system has a baseline that holds actual meaning. There are tons of mags that place dollar value on players quite accurately for ROTO leagues. (and with the value of SBs and Avg. in ROTO, Ichiro is probably doing quite well).
As to the question of whether Ichiro was worth $17 million in 2008. The answer is - he CANNOT be. The team won 61 games. His nearly league average 102 OPS+ placed him right next to Ryan Sweeney - the Oakland RF.
Considering the Mariners 26th ranked team DER, (compared to Oakland's 4th), I'm willing to accept that Ichiro did NOT make a significant positive defensive difference in 2008. Truthfully, Ryan Sweeney only played 115 games in 2008, (but was 9 of 10 in SBs). From a production standpoint, Ichiro WAS Ryan Sweeney - except he did it for an extra 50 games.
If you're trying to convince me that Ichiro was worth $17 million, than you are saying that Ryan Sweeney was worth at least 2/3 of that. Was Ryan Sweeney worth $11 million in 2008? His team won more games. His team played better defense. If you expanded Sweeney's stats to a full year - would ANYONE attempt to state that HE was worth $17 million?
Of course - some might want to bring out the dreaded "intangibles" argument. I say - bring it on. It is my belief that the specific intangibles Ichiro brings to the club likely COSTS them games. I believe his "leadership" has largely been in the wrong direction, and has been part of the problem of Veteran Entitlement (tm) in Seattle. I believe his approach to defense, while probably first rate on a personal level is likely detrimental to his teammates, because he doesn't "look" like he's giving 100% in the field. (Griffey would throw himself all over the place to get to a ball -- Jeter will dive head first into the bleachers -- Ichiro will chop-step and take a single on an easy bounce.)
When Ichiro was ADDED to Boone and Edgar and Olerud, magic happened. Ichiro "became" the team leader and face of the franchise in 2004, after the long-timers folded up and left. But even then it has been Ibanez who has led the team production-wise. The only season Suzuki has led the team in OPS was 2004, when he scored his PERSONAL 262 hit record.
The going rate for a 100 OPS good-glove, no-power OF is about $8 million. That's what Kotsay got in 2008.
There were *FIVE* players in all of baseball who earned more than Ichiro in 2008. AROD, Giambi, Jeter, Manny and Beltran. And no, Jeter wasn't worth his $21 million either.

One addition, Sandy. Ichiro is more marketable to the Seattle area than *most* other cities due to his incredible popularity in Japan, and the strong connection the PNW has with Asia. So in marketing dollars, he's worth more..but I can't even begin to argue that he's worth the ~$9mil you've outlined based entirely on that. A few million, sure. But not the difference.
Good point on the entitlement issue, as well. Hadn't put it in that light before.

I suppose some of my rant is based on my growing irritation with the mass ability to continue viewing an object in the present "as if" it were still 4 years in the past.
Ichiro (2001-2004) OPS+ 126, 120, 112, 130 -- true superstar. You're getting .380+ OBPs from a leadoff hitter, who is stealing 35 a season, and delivering gold glove defense in RF for one of the best defensive teams in baseball.
Ichiro (2005-2008) OPS+ 113, 106, 122, 102 -- nice filler. You're getting .360ish OBPs from a leadoff hitter, who is still stealing 35 a season, and delivering imperceptibly declining defensive play for a team that is one of the worst defensive teams in baseball.
First 4 years -- one season UNDER 120 OPS+.
Next 4 years -- one season ABOVE 120 OPS+.
An objective view of Ichiro's career, and you see a guy who was an .830 OPS leadoff dream, who is now a .780 OPS leadoff good-to-have.
The contention that Ichiro is somehow STILL worthy of being the 5th highest paid player in baseball is ludicrous in the extreme. It's been almost 4 years since he actually WAS that guy.
You know how many OUTFIELDERS had more total bases in 2008 than Ichiro? 27.
If we look at all hitters, Ichiro would be 57th.
Of course, total bases seems like a strange route to go -- except that it should, in fact, BOOST Ichiro's status, since he doesn't walk much. (97 guys had more walks).
Attempting to try and build a case that in 2008, Ichiro actually produced enough that he was the 5th most productive player in the game is beyond laughable. The only universe where one could even raise the question would be one where the ONLY things that mattered were hits and SBs. So, yes, in a 2 X 2 ROTO league, maybe Ichiro could be in the top 5 in value.

sorry taro,
In 2008, 17 million was the 5th highest salary in all of baseball. So, unless Ichiro is the 5th most productive player in baseball, then he cannot be viewed as "clearly" worth 17 million.
In 2008, Ichiro played 90 games in RF, (where above average hitters are plentiful). He only played 69 games in Center.
Your premise, that finding above average defensive CFs with above average offense, is patently false. Let us simply look at the CF production during 2008.
Player -- OPS
Hamilton - .901
Sizemore - .876
Beltran -- .876
Granderson - .859
McLouth - .853
R. Ankiel - .843
V. Wells -- .839
Hunter -- .810
Cameron - .808
C. Ross - .804
Victorino - .799
Kemp ---- .799
Upton -- .784
Schumaker - .765
C. Young .758
Rowand - .749
Ichiro -- .747
=====================
If Ichiro is worth 17 million, then I'm guessing the top 5 on this list are ALL worth more than 25 -- Sizemore, Beltran, McClouth -- these are NOT gloveless wonders shoved into center due to lack of options.
=====================
As to the concept that the system is "FA" dollars, which obivously inflates the numbers -- that IS my point.
If in reality, only 5 players make 17 million -- but in your system, 80 players are "worth" 17 million, then attempting to equate the two becomes utterly useless -- because the economics of reality preclude there actually BEING that many players at those prices.
If the average major leaguer makes $3 million a year -- but the average player in your system is "worth" $9 million a year, the disconnect between system and reality is so vast that any attempt to speak of the two as if they are remotely on the same scale becomes meaningless.
The actual top salaries in baseball are what they are in part BECAUSE of the screwy economic system that creates a 3-year window of minimum play -- and then a slow climb toward actual worth. If every, in reality, was a free agent. If there were no minimum salaries -- and no 3 and 6-year windows to play with, then the entire salary structure would be vastly different. The VERY likely outcome would be the top salaries would be DRASTICALLY lower -- and the entry salaries would be drastically higher -- unless total revenue is changed drastically.