Indians, tech, and Antonetti
You may recall that we’ve mentioned Cleveland’s Chris Antonetti as a possible GM candidate, and in general we’re fans of what that front office has done. Check out this article for some detailed info on the kind of information advantage they’re working with.
It does contain this:
This information, plus another computer analysis that showed no one player’s salary had exceeded 15 percent of a team’s payroll on any World Series champion club since 1985, overrode the Tribe’s emotional instinct to pay Thome the guaranteed salary he wanted for six years to allow him to finish his career in Cleveland.
I can’t believe that that really swayed anyone, it’s such an obvious failure of reasoning. It’s been circulated so widely it’s worn thin, but that doesn’t make it worth anything. I wonder if they spread that because it’s accepted wisdom by many of baseball’s writers and analysts so it can be used to justify unpopular moves, whether or not it’s true.
And yet, there’s a lot about it, including this :
Back in 2000, when the Indians were preparing for negotiations with then-Indians slugger Manny Ramirez, Antonetti examined championship teams’ player salaries. He found that no World Series champion between 1985 and 2000 allocated more than 15 percent of its payroll to a single player. In addition, he determined the higher percentage of payroll a team spent on one player, the lower its winning percentage.
For example, teams that spent 17.5 to 20 percent of their payroll on one player won 47 percent of the time. Teams that spent 7.5 percent or lower on one player won 53 percent of the time.
Antonetti concluded there was a significant decline in a team’s chances to make and advance through the postseason if it allocated more than 15 percent of its payroll to a single player. On average, his analysis found, successful teams spent a little more than 12 percent on their highest paid player.
Not surprisingly, then, Antonetti recommended that Thome’s contract should not exceed 15 percent of the Indians’ team payroll in any season in which management felt the club had a “legitimate” chance to contend for the playoffs. Why? Because they needed the salary flexibility to acquire other players to put together a winning team.
Ideally, Antonetti said, Thome’s salary should make up about 12.5 percent of the payroll.
To refute, briefly: this is the logical fallacy “Cum hoc ergo propter hoc (with this, therefore because of this)”. It’s like discovering the Mariners only contend for division titles when Lucent and Cisco stock is highly valued. Should the Mariners use their huge bank balance to buy up those stocks in an attempt to drive up their price?
Of course not. While “salarly flexibility” sounds good as an explanation (and, I’d argue, roster flexibility is a huge boon to a team) the issue is much simpler than that.
– Teams with high payrolls generally win more.
– High payrolls mean that one player with a big contract does not consume so much of the team’s total salary.
That’s it. Take the Angels, for instance. They have a $95m payroll, and they’re paying their star player Vlad Guerrero $12.5m this year. 13% of payroll.
Now say that he plays on the Royals, or the Devil Rays, or the Pirates. He’d still be the same player, but now he’s make up about 26% of the payroll and the team would be terrible.
It’s the same deal with the Cardinals and Pujols (and, for them, Walker-Rolen-Edmonds).
You can look at examples of teams that have one player who makes a ton of money. They’re bad to awful teams who’ve retained one marquee player, or someone who had a horrible contract they couldn’t dump off on anyone, or even a modest veteran picked up to plug a huge hole in the roster.
Or think about it this way: at a 40m payroll, the threshold for consuming 15% of payroll is only $6m. At the mid-point for teams, it’s about $10-11m. At $90-something, it can be about $13-14m, and at the Yankees’ level, they can have players making $30m without exceeding 15%. It’s not about proportion of payroll consumed at all, and it never has been.
I’m surprised to see someone as smart as Antonetti spending any signficant time researching this.
Still, it’s a good article, and intersting food for thought.
Comments
50 Responses to “Indians, tech, and Antonetti”
Yes, that oft-quoted 15% rule has always bugged me.
Given the background of some of the M’s ownership, it’s really a shame that the Mariners aren’t the leaders in this area. They should be. Back in the late 80s when Larson and Marklyn were developing Microsoft Access, one of its first test uses was to manage their fantasy baseball league.
If you’re going to post, why not say anything?
In any case, regarding Thome and the Cleveland FO, I’d be willing to bet they raised that particular “argument” as a complete smokescreen (which I think is the suggestion Derek was making as well). After all, if they aren’t foolish enough to believe it themselves, there’s no harm in letting others do so . . .
Surely, unless he’s an idiot, he must have controlled for total payroll, right? Obviously you can’t compare the Yankees and the Royals or Devil Rays directly. But what if you just include a group of mid-level teams with similar total salaries? Within say, plus or minus 15% from the average?
I thought for example that the Ranger’s experience with A-rod was a pretty good illustrative example of why you should not put too much eggs in one basket.
Of course, the real key is not to overpay and be stuck with an over 15% guy who doesn’t pull his weight.
Another point against the percentage argument: Every team pays the same league minimum for “freely available, cheap talent.” The Yankees and Royals pay the very same amount to fill out their benches and bullpens. In percentage terms, that’s chump change for the Yanks, but a significant piece of payroll for the Royals.
One can make a similar failure in logic by stating “No team has ever won a championship with a budget below X.”
“He found that no World Series champion between 1985 and 2000 allocated more than 15 percent of its payroll to a single player.”
If Antonetti’s findings are accurately reported here, they are restricted to winning teams and would exclude Tampa and other also-rans. The facts hold this to be true: not a single WS winner back as far as 10 years is out of the top third in payroll, which confirms Derek’s argument in spades.
If you eliminate teams to the extent that you consider only those with WS rings in your sample population, you end up with serious analysis problems.
I know the analysis is presented as having to dug deeper into winning percentages as they relate to single-player payroll imbalance, but the whole is based on a fallacious premise: logical fallacy.
I say give two players 25% or more of your salary, and you can be successful – so long as they are the right two players – e.g., what the Giants did with Bonds and Schmidt.
The Rangers get a lot of credit for dumping A-Rod’s salary as if this was the reason the Rangers on a whole sucked. But the problem was not A-Rod but guys like Chan Ho Park. If Park had been the stud his contract suggested he was, history suggests the Rangers would have been at least .500 in those years. It’s very difficult to find teams with these two ingredients playing sub .500 ball (although someone pointed out to me once that the Killebrew-Pascual Senators of the late 50’s managed to).
The Angels in 2003 followed my own formula for success: If you are going to spend money on payroll do two things first, and let the rest of what you do take care of itself: Get a stud hitter, someone who will give you an OPS of over 1.00. And get yourself a stud starter, a guy that eats a ton of innings effectively. Spend whatever it takes. Then fill your team with whatever you have left – replacement value, rookies, castoffs, guys freely available who can adequately fill a major league role.
The Angels did this when they bought Colon and Guerrero, and let teams like the Mariners outbid them for the services of the Scott Speizios they no longer had room for.
There aren’t a lot of these studs available, but it seems that at least one is on the market every year. Do what it takes to get one of each, and work from there.
The Indians didn’t screw up (long term) by not paying Thome – because Hafner has filled in very nicely thank you. But they probably should have kept Colon.
Now, for a lot of teams, .500 is not exactly what they are shooting for. But if you are small market, or big market inept like the M’s – it is a pretty good strategy. Once you are at .500 – then you are a few good/lucky moves and a suprize career year or two away from contention. Or maybe, if you are the Padres, you are there already.
But they probably should have kept Colon.
No. This isn’t really open for debate. Do you remember what they got for trading Colon?
Cliff Lee, Brandon Phillips, and Grady Freaking Sizemore.
Sizemore alone was worth as much to the Indians this year as Colon to the Angels. Add Cliff Lee, and… man. Even considering Brandon Phillips hasn’t lived up to expectations, Colon was a FA-to-be who would (and did) command far too much for far too long on the open market.
And they’re paying those guys $5 and a ham sandwitch a day.
Doesn’t the 15% argument also make another error in logic? It seems like the fallacy of arguing to tradition…but using numbers to do it.
In addition, he determined the higher percentage of payroll a team spent on one player, the lower its winning percentage.
Well of course. Given any player salariy, the larger the percentage of the whole team salary the lower the team salary.
So he basically said that teams with higher payrolls tend to win more. The Yankees alone give us a positive r-squared on that one.
It is interesting to note that since 2000 every world series champ has spent at least 15% of their total on a single player: ’04 BoSox/Manny (the player that Antonetti decided to let go in 2000 because of the 15% theory), ’03 Fla/Ivan Rodriguez, ’02 Ana/Salmon & Appier, ’01 Az/Johnsnon.
Also #6 the ’03 Marlins are the exception, their total salary was in the bottom third in 03. I believe they also spent the highest percentage on a single player since 1985, over 20%, on a single player.
If the Astros win the series they will destroy Antonetti’s theory with two players that each make up 23% of the teams total salary on their own (Bagwell and Clemens at $18M each). Although none of the other teams remaining have a player over 15%.
Oooh! Oooh!
I wonder what DiamondView would spit out for Darrin “Gritty” Erstad and Willie “Dynamite” Bloomquist…
I’m quite sure, though, they’ve got DV quite tweaked to put those intangibles into proper perspective…
Having seen DiamondView in person and looking at its capabilities, I can vouch that it frekaing rocks.
N°9 (could resist the Lennon reference): Trading Colon for Lee, Phillips and SIZEMORE would have been a bargain, but my understanding is that Colon signed with The AAFLA as a free agent. That from Retro.
Colon did sign with the Angels as a free agent. He was traded to Montreal for Lee, Phillips, and Sizemore.
Derek is absolutely correct that winning depends more on how much you spend than on how you allocate spending among players. However, it does not follow that careful analysis of the allocation issue should be dismissed. In fact, I contend that it is a vital extension of the basic value issues expounded in “Moneyball†and I would be surprised only if someone as smart as Antonetti was NOT “spending any significant time researching this.â€Â
As Derek mentions above, “teams with high payrolls generally win more.†Beyond that, the value issue – whether actual performance closely corresponds to pay – also is a significant factor. Witness the ’05 Yankees, who at $206 million made the playoffs by only two games over Cleveland, at $46 million. Total spending and value are two factors whose effect on winning are widely acknowledged and understood.
The third issue, which Cleveland was addressing with Thome, is allocation of payroll among roster spots. Analytically speaking it is a classic optimization problem, the kind you’d solve with a large matrix, and frame something like “Maximize expected VORP, subject to the market conditions for players at each position and batting order spot.†I find it quite plausible that the matrix cannot be optimized if one cell exceeds 15% of the total, especially if the free agent market is characterized by diminishing returns in performance as annual salary approaches and exceeds $10 million per year.
Sure, it’s easy to find logical fault with the generalized and simplified terms in which the Indians’ decision is proferred to and promulgated by the mainstream media. But given Antonetti’s education and aptitude, high-level optimization analysis is screaming from between the lines of the article. No proponent of an advanced analytical approach to building a winning baseball team should dismiss the “15% rule†lightly.
I don’t understand why the comments here are so critical of the 15% rule (I’d think of it more as a “guideline”). It strikes me as a pretty logical approach for spending money in a “team” sport. Baseball is not like basketball where you can put the ball in the hands of your best player when things are on the line. With the exception of bringing a stud relief pitcher in at critical moments, you are pretty much forced to play averages, and it seems to me that does argue in favor of avoiding a payroll that is overly weighted to one guy. Of course this means a team with a lower payroll can’t afford to pay a Vlad or a Thome the same amount as a high payroll team, but that makes sense to me. I think it’s a poor percentage strategy for a low payroll team to think they’ll assemble a winner by spending their wad on a single player. The low payroll team probably can get more marginal benefit from those dollars spreading it around. Of course there will be exceptions to every rule (or guideline), but I like Beane’s approach of developing talent, offering mid-level money to keep the best of that talent for a while, and let the high priced guys leave (in exchange for prospects or draft picks).
This is not unlike the “no team with a 50-homer hitter has won the WS” rule you used to hear so much about.
jason
how about the no-Cub player rule?
It would seem that the more correct question to ask is: “What does it take to get championship-caliber performance out of a middle-of-the-road budget?” And the immediate answer is: “You have to get more performance out of your players than what you’re paying them for.” If you pay middle-of-the-road money and get average performance for that money, you’ll have a middle-of-the-road team, end of discussion.
So if the argument is made that paying 15% of budget to Jim Thome or Bartolo Colon would hurt a team, the natural follow-up question is: How does it hurt the amount of performance you get for the money? It can imaginably do so in either of two ways: either by the big-money guy not producing in proportion to his pay, or by the big-money guy’s presence undermining the performance of other players (this is the “chemistry” argument).
But the 15% rule is often stated so as to imply that your championship chances would be less if you have a $95 million team with Jim Thome getting $15M, than if you have an $81M team with $1M going to Ben Broussard in his place (with the other $80M being shared identically in either case by the other 24 guys, none of whom get more than $12M). If this version of the 15% rule is used, then it requires a “chemistry” argument. Nobody would claim that Thome’s likely performance would be worse than Broussard’s (nor A-Rod’s worse than Carlos Guillen’s, etc.), so they must be implying that Thome suddenly getting a big contract would make the other guys play worse– indeed, more than five games worse (to make up for the 50-run difference between Thome’s 2003 VORP and Broussard’s).
This assertion can be tested, although not easily (translation: I’m not going to do it here), and certainly not by looking at 15 years of World Series winners. It would take something more like a study of PECOTA predictions vs. actual performance for a large group of teams with more than 15% of salary concentrated in one player. The World Series observation is too easily explained by the facts that (1) (as others have mentioned above) that the teams that can be most realistically forecast to win big are those that can afford multiple stars; and that (2) although some championships are won by no-star teams for whom everything comes together in one year, the fact that these teams outnumber those with one star at 15% makes it more likely that a given year’s most overperforming team will be a no-star rather than a one-star team, even if the likelihood of success for any given one-star team isn’t any worse.
Wouldn’t it be simpler to say you can’t put all your eggs in one basket? And that you’re not going to get a championship if the surrounding players are way below average (i.e., not going to get there with average performers or with high performing rookies, so you have to pay to get that above average performance throughout that lineup).
Derek is absolutely correct that winning depends more on how much you spend than on how you allocate spending among players. However, it does not follow that careful analysis of the allocation issue should be dismissed.
This is the straw man fallacy. I never said that, I’ve written about the resource allocation issue frequently.
In fact, I contend that it is a vital extension of the basic value issues expounded in “Moneyball†and I would be surprised only if someone as smart as Antonetti was NOT “spending any significant time researching this.â€Â
Again: it’s not the issue of resource allocation. It’s that he was looking at the 15% thing, which is silly.
[snip]
[snip another straw man]
Sure, it’s easy to find logical fault with the generalized and simplified terms in which the Indians’ decision is proferred to and promulgated by the mainstream media.
That you assume that this is the case is not borne out by the evidence. This is a clear case where they did a lot of work on a hare-brained theory.
But given Antonetti’s education and aptitude, high-level optimization analysis is screaming from between the lines of the article.
I’m certain they do this kind of thinking. It’s not in the article.
No proponent of an advanced analytical approach to building a winning baseball team should dismiss the “15% rule†lightly.
Well, gosh, if you say so.
It’s a bunk rule. It’s demolished with any sort of logical thought or research. This is like arguing that anyone who’s serious about learning about cosmology has to read the astrology column every day. It’s junk.
Also, Paul’s a Super Reader.
To me, the interesting question is in regards to the distribution of talent on a team. Does a team with an even distribution of talent win more or does a team with “stars and scrubs” win more. Once that question is answered, at least the team can determine whether it makes sense to substitute two good players for a star and scrub. I don’t think the analysis of this would be too difficult (in theory at least). All you would need to do is control for the overall level of talent of each team and isolate the effect of the distribution of talent. If this question is answered through, for example, a regression of wins on team VORP and some metric of the distribution of VORP (gini coefficient maybe), at least teams will know what to strive for. If even distributions tend to win more (after controlling for the overall talent level), teams should tend to avoid the “stars and scrubs” approach in favor of the type of team the M’s won with a few years ago in which there were many very good players but no real stars. If uneven distributions tend to win more, teams should pay for better players (the stars) by cutting salary in other spots (by hiring scrubs). I have a feeling Antonetti was basing his overly simplified statement on analysis similar to this rather than some stupid observation that no team has won with more than 15% given to one player. I just have trouble believing someone in his position would take such simplified analysis and treat it as if it meant something when, for reasons previously pointed out by others, it clearly doesn’t.
A complication in the above analysis is that teams’ desire to formulate their team as an even or uneven distribution of talent is certainly affected by the price of talent. For example, if “stars and scrubs” turns out to be the better distribution of talent, the price of stars will reflect that and salaries will increase at an increasing rate with respect to VORP (or whatever metric you prefer to measure talent). If even distributions tend to win more, there will be a greater demand for mid-level talent and player salary would be more linearly related to VORP or maybe even increase at a decreasing rate. So, I believe that the optimal distribution of talent clearly depends on how talent is priced. If, for example, the general sentiment is that “stars and scrubs” is the right approach, there may be value in formulating a team around mid-level talent because, given the feelings of others, this talent will be undervalued. This is the idea behind Moneyball if I’m not mistaken. Find the talent that other people undervalue and acquire it.
Is there any chance the article is misunderstanding what he was studying? Perhaps it was more along the lines of the corellation of one’s percentage of payroll to one’s percentage of expected production?
Regardless, studying 15 World Series’ teams is too small a sample to draw any useful conclusions anyway. And why would you study W.S. teams instead of the teams with the best records?
Both good points, adroit.
Mfan: Nate Silver wrote a really nice piece about the cost of distributing talent in a recent Baseball Prospectus article that talks about some of the issues you raise.
DMZ – Do recall what it was titled? I am interested in reading it, but in a quick glance through the Nate Silver archive I didn’t see a title that would match with the topic of interest.
“Mulligan on Guzman“. Starts off talking about something else, but you’ll see the stuff that interests you pretty quick.
The biggest problem with the 15% rule is that it means that a contending team that obeys the rule necessarily gets worse simply by adding one really good player at price exceeding 15%.
And that’s absurd.
Oakland had a good team worth $50 million. There’s no way adding A-Rod for $25 million (thus using 33% of the payroll on one guy) would have made the team instantly less good.
Evan: that is exactly right. Except that Beane would neer hire A-Rod.
In reference to N°16: the subject was Cleveland-to-Anaheim, not Toronto.
Cleveland didn’t deal with Anaheim, though. Cleveland traded Colon to Montreal for Lee, Phillips, and Sizemore. Montreal subsequently traded Colon to the White Sox for El Duque. And then Colon became a free agent, and signed with Anahiem.
Read Rick again.
I don’t disagree with anything you’ve said, I only said that you agreed to change the goalposts.
Toronto also got David Wells for Clemens, but that also has nothing to do with…
The point was made that Cleveland should have kept Colon. Given who they got for him, that was disputed.
Then you came along and said, “Trading Colon for Lee, Phillips and SIZEMORE would have been a bargain, but my understanding is that Colon signed with AAFLA as a free agent.”
Colon was traded for Lee, Phillips and Sizemore. Colon did sign with Anaheim as a free agent.
I don’t see how the two are connected at all.
Here is the simplest way to debunk the nonsense about the “15% rule”
What is magic about 15%, why not 14% or 16%? Unless the proponents of the rule can show some mathematical reason based on roster sizes and allocation of talent or something it is nothing but an arbitrary number that coincidentally lines up with a recent series of world series winners.
Or put another way if the world series winner in ’92 had a player making 16.3% of payroll then the rule would be changed to the “16.3% rule”
Paul Covert in 21 and Evan in 30 are far more egregious purveyors of “straw man” argumentation than me, with their argumentum ad absurdio applications of the “15% rule.” Derek, for his part, pretty much ignored the meat of my position in his [snip] [snip] snippy reply. So allow me to reiterate, from a slightly different perspective.
The “15% rule†is not a hard-and-fast rule in and of itself. Of course any team would be better each time they added a star player at the expense of a scrub and a 17.65% increase in payroll. However, that is not the problem the Indians, or any baseball team, is faced with. They are given with a payroll budget, and have to decide how best to spend it. It’s payroll allocation, and like I said above, that’s a classic optimization problem. Antonetti is not claiming you WON’T or CAN’T be successful if you spend 15% on one guy, simply that you’re LESS LIKELY to be successful. It’s not optimal use of resources. Do teams win pennants with less than optimal use of resources? Sure, all the time. Should you ignore optimal use, especially on a payroll as limited as Cleveland’s? Not if you want to be a winner.
Now, of course, I’m placing a lot of faith in Antonetti’s abilities, because I’m familiar with the type of business analysis he’s doing. That, plus David says his DiamondView “freaking rocks.†If Dusty Baker says he’d rather have three talented, hustling youngsters than one Sammy Sosa, I’ll take that with a grain of salt. But when a bright young stathead like Antonetti tells me something along the same lines, I’m going to sit up and pay attention.
The Indians look brilliant right now, but the cost was high. They were bad for a couple of years and lost a good deal of fan base during that time. Had they kept Colon, I’m suggesting they could have continued to rebuild AND may have found a way to stay in contention – give their fans something to hope for each year, particularly in what was pretty much a weak division.
I’m suggesting that an organization that scouts and develops young talent like Cleveland does, AND has a shiny new ballpark and a large fan base, shouldn’t necessarily feel they have to piss away the fan base for a number of years in order to rebuild the way they did. And a small market team shouldn’t feel like they have to suck for a decade in order to get lucky with the right combo of young players who haven’t yet hit the free agent market.
I can’t remember all the particulars of the Indians contractual obligations – perhaps they had no choice – but I would say that a team would do well to keep any pitcher and position player who each log in a large number of innings with “top 10” MLB success. Keep one of each, and build around them with what you have left. You’ll do a better job of balancing short term/long term needs.
Evan: “Trading Colon for Lee…” was in reference to what Cleveland could have gotton for Colon. And Cleveland got exactly nothing. You didn’t change the subject but only chose the follow the thread of someone of authority who had.
The trade of Colon for Sizemore & Co. is a valid argument if talking about Toronto, but Toronto was never in the mix. All tha approach does it to set some fairly ancient trade value for Colon with a team that was owned by MLB and that no longer exists in a form resembling what it was in 2002.
So I have your attention. Yes: Toronto still exists and was not an MLB team. Please look at comments 9 and 10. Writing from Paris means I’m out for now.
Have fun with the game thread tonight.
Brian: Yes, your version of the rule is much better– certainly reasonably arguable, whether it’s correct or not. If indeed Antonetti was being serious (rather than smokescreening to justify an intelligent but unpopular decision to the public), then your way is certainly how he must have meant it. But given that we live in a world full of player-turned-broadcasters and fans who take them seriously, I certainly don’t think the crude version is a strawman argument.
In the more sophisticated version of the 15% rule (treating it as an approximation to a sliding scale, as you describe), the question becomes whether it’s better to spend (say) $15M on one guy, or $7.5M on each of two guys (or something similar). The argument for going after the star is that, for positions without a veteran-contract commitment, a good organization should be able to find a bargain player that’s being ignored by other teams, or from its own farm system; and that therefore there is efficiency in having your veteran contracts concentrated into a relatively small number of positions (unless you have Steinbrenner’s budget, of course– but then you don’t have to worry about efficiency).
More precisely, this argument justifies at least some amount of premium (measured in Marginal$/MarginalW) for the major stars. I’m not claiming that going for the superstars is always and inherently the correct decision, mind you; if the superstar premium on the market got too high, the correct policy could reverse. But even then, it would be a “star vs. midranges, given available budget” question, not a “percentage of payroll” question.
In A-Rod’s case, for example, I do happen to think that Texas gave him too high a premium; just not as drastically so as it was typically made out to be. In Thome’s case, it was indeed a good decision not to spend big for him; but not because of the percentage-of-payroll argument. Rather, it was a good decision because he was an over-30 first baseman with a high risk of falling short of his late-20’s performance level (the Branch Rickey argument).
In either case, though, the choice of stars vs. midrange guys won’t be the driving factor in whether you can get a championship-quality team on a midrange budget. The key won’t be the guys whom you pay for a predictable performance; the key will be getting guys who can contribute well in their first six (especially their first three) years, plus guys whom the market overlooks.
In either case, this is a matter of out-scouting the other teams (whether in the amateur market or among the low-range free agents). It’s what Billy Beane is famous for, although Billy uses it for a different purpose (he uses it to form an above-average team on a bare-bones budget, rather than a champion on a midrange budget; but the strategy is the same). But for a team that can’t spot an underpriced or underdrafted player (not that we know of any such teams *coughcough*), there’s not an acquisition strategy in the world that will make them a champion.
Paul Covert in 21 and Evan in 30 are far more egregious purveyors of “straw man†argumentation than me, with their argumentum ad absurdio applications of the “15% rule.†Derek, for his part, pretty much ignored the meat of my position in his [snip] [snip] snippy reply. So allow me to reiterate, from a slightly different perspective.
Wow, it’s the “you too” falcity dragged in. Yayyy.
I’m sorry if you disagree with my choice of editing. I chose to argue in particular with your assertions that tossing aside the 15% rule was somehow a vote for a march back into the Stone Age of analytical thinking, and I made my displeasure the point of my reply. That my reply was not in agreement with you, or pleasing to you in tone or composition, is unfortunate.
For everything else, Paul’s got it.
[i]Evan: “Trading Colon for Lee…†was in reference to what Cleveland could have gotton for Colon. And Cleveland got exactly nothing. You didn’t change the subject but only chose the follow the thread of someone of authority who had.
The trade of Colon for Sizemore & Co. is a valid argument if talking about Toronto, but Toronto was never in the mix. All tha approach does it to set some fairly ancient trade value for Colon with a team that was owned by MLB and that no longer exists in a form resembling what it was in 2002. [/i]
*blink* How did Cleveland get nothing… ? The comparison to what Colon is worth to the Angels and what Cleveland got for him from [b]Montreal[/b] is valid. Bartolo Colon’s market value was not created in Montreal or Chiacgo, and one could reasonably argue had Cleveland not traded him he would have signed elsewhere for a similar amount after Cleveland refused to re-sign him. What Cleveland got for Colon is paying off NOW (and those players are proving to be a better investment for the money) than keeping Colon would have been. Tell me again how Cleveland got nothing for Colon?
And what the heck does Toronto have to do with Bartolo Colon?
Apparently [i] and [b] don’t work. Hooray for reading the small print 🙂
Soooooooooooo after trimming this back a little to get on topic — it’s interesting to me that only a few people were interested in the data advantage. There are a lot of teams that have designed and built their own defensive metrics — the A’s, I know, have one, the Red Sox do a lot of research into player evaluation and prediction, which is part of why they hired Voros.
But we don’t hear anything about the M’s doing this. They look to Olkin as a kind of… outside voice, I think I’d say. In a sense, it’s a style thing: they’re an old-school baseball organization, even as I say that that’s not, by itself, good or bad, but it does (as someone else noted) strike me as strange that the Mariners, who count among their minority owners some of the best and brightest of the tech boomers in Seattle, don’t have some kind of baseball lab downtown where guys wearing 90s-style Buhner and Griffey jerseys drink coffee and argue about the best way to compensate for foul ground in equalizing park effects.
It would seem like a natural fit, and a place where the team’s location and resources could help it win games.
DMZ, from what you can tell, are the Mariners even using Olkin for anything useful? It seems his role is very reactive — Bavasi sends him a specific question and he comes back with an answer, e.g. we’re thinking about trading Joel Pineiro for Derek Lowe, tell us what the stats show on Lowe; “yuck”. Whereas he could be proactive, looking into general projects or ones that interest him — what kind of hitters would have an Ibanez-like edge in Safeco, how might the Mariners make large improvements for 2006 cheaply, a risk-benefit analysis of all of the free agent pitchers alongside of Moyer, Meche, and Franklin.
In some sense they could get just as useful consulting by adopting an alias and posting on this blog.
Another data fact that the Indians could have thrown at Thome is that from 1985 to 2001, no team had won the World Series with a leader in HR’s that hit over 36. Rob Neyer pointed this out as the Curse of the Balboni, as Steve Balboni hit 36 HR’s to lead the KC team which won the WS.
In 2001 Arizona and Luis Gonzalez finally broke this curse but it does show that if you’re going to use a sample size as small as say 16 or 17 years and use the sample to define a key characteristic of the winning World Series team, there are all kinds of possibilities.
DMZ, from what you can tell, are the Mariners even using Olkin for anything useful?
Sure.
The implied question in your comment is “are they using him as well as they could?” and this is a much more difficult question to answer than it might initially appear.
They brought in Olkin to do what you imply: consult on decisions. I think Olking also has some brainstorms about certain players or moves that he passes on. This is what the team wants and, for lack of a better term, will take from a stat consultant.
Now, if it was me, I’d have a crack team of Olkins working on arbitration cases, constructing a super data store like the Indians have, except 20% more awesome, crunching random stats to find players worth sending scouts out to investigate, minor league free agents to hire, and people working on evaluating player evaluation and projection tools — everything where you can apply performance evaluation to winning or running the baseball side more efficently.
However, I’m not Bavasi. If he and the other front office guys feel they’re better off getting that information in different ways, and applying Olkin to certain particular tasks that come up, well… that’s how they’ve decided they can best be helped. Having a crew of baseball brains they don’t have the organizational energy to listen to wouldn’t help anyone.
So given the nature of the Mariner organization and the people who compose it, I’m happy to know that Olkin’s there and consulting on decisions. It’s more than many teams of old-school baseball types have. I hope it’s a successful partnership, and that it leads to the team trusting him with more and more questions and consultations.
This is a general apology to everyone and to Evan and Rick in particular for my rambling, stupid and inaccurate comments last night on this thread. I started taking heavy doses of muscle relaxants, pain killers and anti-inflammatory drugs yesterday for a slipped disc and the combination clearly got the best of me. It won’t happen again.
Sorry.
Now, if it was me, I’d have a crack team of Olkins working on arbitration cases…
This is the true selling point for statistical analysis, improving the bottom line. I’ve been part of organizations that have tried to sell various cool products and services, and you always get a meeting with a potential customer because they want to see what’s cool, cutting edge, and state of the art. But when it comes to buying, they show you the door until you show how it can improve the bottom line. And if your idea increases share price directly that’s platinum, increase revenue – gold, and cuts costs – silver.
So to be more specific, if you constructed a data store that helped to cut total arbitration costs by 50% while maintaining or improving current productivity levels for players with 3-6 years of major league experience on your roster, that is something management is likely to listen to.
Expected wins above replacement x 2 = millions of dollars in salary
The average team (81 wins) will be about 35 wins above replacement, or have a 70 million$ payroll. A top team (100 wins) should have a payroll of 108 million$.
If the player in question has a true talent level of 9 wins above replacement, he deserves 18 million$. That’s 16.7% (18/108) of the top team payroll and 25.7% (18/70) of the average team payroll.
The only thing a team can play with is the marginal dollar value per win.
The “2 million$” guideline is what should be remembered.