Selective Application of Statistics

[200tang]

Wow, an interesting Christmas on the Forum indeed. Great to see people act so grown up and being so nice to each other. It looks like the Christmas spirit or even common courtesy has become a "strawman." as well...

Only like 4 people on the forums celebrate Christmas.

[HoustonGM]

however....those who use these stats should be careful not to discredit the obvious. Not to discredit good scouting and in some cases the simple "eye test". There are IMO alot of things that these newage methods can't yet factor yet try to unsuccessfully. I'm learning that defense is definetly one of them. I admit, normally its unfair for me to make that claim not knowing all that goes into their equation, but my eye test in this situation is more than enough for me to make that claim IMO. I simply can never buy the claim that factoring in all factors (bay 09 LF performance in Boston vs. cameron 09 CF performance in MIL) that Cameron was worth 20+% more in terms of dollars than Bay. Thats a fail.

For one, the "eye test" for me matches up just fine with what the stats...well, FanGraphs WAR...says.

For two, can you buy the claim that Cameron was worth 8 more runs than Bay over the course of the entire season, or about 1 run every 3 weeks?

After all, I hope history would tell them not to be so foolish as to believe their newage stats won't one day be in the "previously accepted" category as well. Scouting, common sense, and a keen eye have been IMO the only measurement that has remained constant throughout baseballs history. Even in this newage statistical period, teams still have scouts that watch major league and minor league players. If they were irrelevant and teams TRULY BELIEVED the numbers were 100% accurate they'd simply save tons of money by firing them all and using only computers.

Who has ever said that scouting and such is irrelevant? This is another all too common yet completely absurd strawman.

[200tang]

Who has ever said that scouting and such is irrelevant? This is another all too common yet completely absurd strawman.

Pretty much. I don't think I've ever seen anyone say stats are always better than scouting. This is just a common argument by people who don't know what they're talking about. It's always about a blend of stats & scouting.

[HoustonGM]

Also, looking at WAR in terms of percentage is very misleading. A 4 win player, about an All Star caliber player, would thus be "100%" better than an average player (roughly 2 wins). Pretty misleading

[kenny1234]

Tom Tango is currently one of the leading sabermetricians in the game, but that doesn't prevent him from enlisting the support of fans of all 30 teams every year to help compile his annual "Scouting Report For The Fans By The Fans".

I generally agree with your post - but I don't think that the Fan's Scouting Reports, like Fan Projections, are anything but a reflection of the statistical data. The people that post on that site are people that follow the statistical data - so asking them to define who are good players and who aren't is not an unbiased check on the accuracy of statistical measures of defensive ability. The two coincide because the fans that post know the statistical evidence, and simply parrot it back.

[kenny1234]

Also, looking at WAR in terms of percentage is very misleading. A 4 win player, about an All Star caliber player, would thus be "100%" better than an average player (roughly 2 wins). Pretty misleading

If the translation of WAR to a dollar figure is meaningful, then looking at WAR in terms of percentages is meaningful. It might be reasonable to say that neither makes sense, but if someone says that Cameron is worth $10m and Bay is worth $8m, then it isn't wrong to say that Cameron is 25% better. WAR defines the metric with which you are measuring players.

[gRYFYN1]

while I feel this may be off the topic of complaining about each other i found this an intersting tool

http://www.baseballmusings.com/cgi-bin/RBIPCT.py

uses retro sheet data to calculate the % of runners available driven in by a player.

[dickay]

You mention the "eye test". Just like Mark Twain's saying about lies, damned lies and statistics, our eyes can fool us as well.

I absolutely agree with this, and have mentioned as such. Thats why I appreciate stats as they can and have changed my mind on things. I don't however see the same appreciation in the contrary. Whenever disagreeing with a stat with a "stathead" their views are always absolute because the numbers say so. In my experiences, and I'll say that my opinion exists in these forums as well, its nothing but a "strawman", as you call, it to say a statitician buys into the eye test, or to be more just...gives it any consideration whatsoever.

For two, can you buy the claim that Cameron was worth 8 more runs than Bay over the course of the entire season, or about 1 run every 3 weeks?

Absolutely not.

Who has ever said that scouting and such is irrelevant? This is another all too common yet completely absurd strawman.

Pretty much. I don't think I've ever seen anyone say stats are always better than scouting. This is just a common argument by people who don't know what they're talking about. It's always about a blend of stats & scouting.

These two comments are just spin city and the type of arrogance I was referring to. Others opinions are cast off because they are apparently "ignorant" in the minds of statheads, ie. "people who don't know what they're talking about". If its truly a blend of stats and scouting than one would have to agree IMO that the stats are flawed because no scout IMO will say that the 09 Cameron was worth more than the 09 Bay.

[actionjackson]

I generally agree with your post - but I don't think that the Fan's Scouting Reports, like Fan Projections, are anything but a reflection of the statistical data. The people that post on that site are people that follow the statistical data - so asking them to define who are good players and who aren't is not an unbiased check on the accuracy of statistical measures of defensive ability. The two coincide because the fans that post know the statistical evidence, and simply parrot it back.

Agreed. Since we're still somewhat on the Bay/Cameron thing, lemme just mash three of the different types of Wins Above Replacement measures together and see if things improve.

CHONE WAR:

Bay: 5.2
Cameron: 3.6

fangraphs WAR:

Bay: 3.5
Cameron: 4.3

Baseball Prospectus' WARP1:

Bay: 4.0
Cameron: 4.1

AVG

Bay: 12.7/3 = 4.23
Cameron: 12.0/3 = 4.00

So, there you have it dickay, Bay was slightly better than Cameron last year when we mash all these value metrics together. I'm not sure how valid what I just did is, but who gives a s**t, I'm running with it. I take it you're a Red Sox fan from your screen name and that explains your concern with the Bay/Cameron comparison. My question for you is:

If you have two outfielders who offer differing skill sets, but who are relatively close in value, who do you take if you're Theo Epstein? The guy who you can sign for two years at $7.25 mil per plus a $1 mil signing bonus, but whose contract will take you through his age 38 season, or the guy who spurned your 4 year $60 mil offer last July and is now finding it tough to get a gig because he perhaps overestimated his value? Bay would be with them through 2013 (age 34) if he signed a four year deal. Why sign an OF to a contract that's twice as long and pays twice as much per year, when the player doesn't offer significantly more value? It might be worth it to sign him, so he could slide into the DH role come 2011, but probably not at the price he's asking. He's probably more valuable as a DH, so that might make sense, but then 2010 becomes a bit messy in the OF while Big Papi is on his farewell tour at DH. Of course there are always injuries to consider as well.

Another thing to look at is their farm system and it's loaded with OF prospects. Josh Reddick is knocking on the door. Ryan Kalish is probably a year or so away and further away than that are Ryan Westmoreland and further still Reymond Fuentes. That's a lot of prospects and they're all graded pretty highly by the various people that keep track of these things. No prospect is a sure thing, but these guys are all very highly regarded by those in the know. Throw in Ellsbury as an MLB starter and Hermida as an MLB backup and things don't look too bad in the OF once Drew and Cameron's contracts come off the books after '11. There's not much point in clogging things up with a big contract that runs for 4 years when you can have a couple of guys for two years and allow the studs a chance to break through, while also offering them time to develop by having quality veterans at the big league level. Drew could probably replace Ortiz at DH in '11 if Reddick shows he can handle things this year.

[dickay]

Agreed. Since we're still somewhat on the Bay/Cameron thing, lemme just mash three of the different types of Wins Above Replacement measures together and see if things improve.

CHONE WAR:

Bay: 5.2
Cameron: 3.6

fangraphs WAR:

Bay: 3.5
Cameron: 4.3

Baseball Prospectus' WARP1:

Bay: 4.0
Cameron: 4.1

AVG

Bay: 12.7/3 = 4.23
Cameron: 12.0/3 = 4.00

So, there you have it dickay, Bay was slightly better than Cameron last year when we mash all these value metrics together. I'm not sure how valid what I just did is, but who gives a s**t, I'm running with it. I take it you're a Red Sox fan from your screen name and that explains your concern with the Bay/Cameron comparison. My question for you is:

If you have two outfielders who offer differing skill sets, but who are relatively close in value, who do you take if you're Theo Epstein? The guy who you can sign for two years at $7.25 mil per plus a $1 mil signing bonus, but whose contract will take you through his age 38 season, or the guy who spurned your 4 year $60 mil offer last July and is now finding it tough to get a gig because he perhaps overestimated his value? Bay would be with them through 2013 (age 34) if he signed a four year deal. Why sign an OF to a contract that's twice as long and pays twice as much per year, when the player doesn't offer significantly more value? It might be worth it to sign him, so he could slide into the DH role come 2011, but probably not at the price he's asking. He's probably more valuable as a DH, so that might make sense, but then 2010 becomes a bit messy in the OF while Big Papi is on his farewell tour at DH. Of course there are always injuries to consider as well.

Another thing to look at is their farm system and it's loaded with OF prospects. Josh Reddick is knocking on the door. Ryan Kalish is probably a year or so away and further away than that are Ryan Westmoreland and further still Reymond Fuentes. That's a lot of prospects and they're all graded pretty highly by the various people that keep track of these things. No prospect is a sure thing, but these guys are all very highly regarded by those in the know. Throw in Ellsbury as an MLB starter and Hermida as an MLB backup and things don't look too bad in the OF once Drew and Cameron's contracts come off the books after '11. There's not much point in clogging things up with a big contract that runs for 4 years when you can have a couple of guys for two years and allow the studs a chance to break through, while also offering them time to develop by having quality veterans at the big league level. Drew could probably replace Ortiz at DH in '11 if Reddick shows he can handle things this year.

i have zero problem with them not signing bay...in fact I hope they don't. i take the homer hat off in this comparison, and am merely pointing out that I don't see how one system can say Bay was even with Cameron let alone worse by quite a bit regardless of how some want to spin it.

The better question is, why are the three reviews so different..or at least why is fangraphs significantly different?

[actionjackson]

i have zero problem with them not signing bay...in fact I hope they don't. i take the homer hat off in this comparison, and am merely pointing out that I don't see how one system can say Bay was even with Cameron let alone worse by quite a bit regardless of how some want to spin it.

The better question is, why are the three reviews so different..or at least why is fangraphs significantly different?

Differing formulas make the results different. One day we'll get a handle on it, but right now measuring defense is such a new frontier, that expecting them to get it "right" is rather futile. It'll evolve and it's certainly a quantum leap beyond PO, A, E, FPCT etc. Hopefully the days of rewarding the slug who doesn't get to many balls, but makes the play when he does whilst punishing the waterbug who gets to so many more balls and therefore flubs plays more often, will eventually be behind us. I think it's kind of cool that this stuff is even being attempted and I look forward to where it'll go next.

[dickay]

Differing formulas make the results different. One day we'll get a handle on it, but right now measuring defense is such a new frontier, that expecting them to get it "right" is rather futile. It'll evolve and it's certainly a quantum leap beyond PO, A, E, FPCT etc. Hopefully the days of rewarding the slug who doesn't get to many balls, but makes the play when he does whilst punishing the waterbug who gets to so many more balls and therefore flubs plays more often, will eventually be behind us. I think it's kind of cool that this stuff is even being attempted and I look forward to where it'll go next.

i agree with this...and think the older stats say very little about defensive performance. I honestly don't know or care to know the intrinquicies of these new methods at the moment. For the most part, I usually buy their interpretations of who is the better defender. What I have trouble with is how they include defense into the grand scheme of things and determine how player A's great defense makes him more valuable than player B and his great offense. While in alot of cases I may feel that is true, I personally don't feel it is 100% as is obvious with my opinions on the Cameron matter. I think Bay's total package was far more valuable than Camerons contributions, especially considering that I don't think Bay was that bad a defender last year from what I saw, and I don't value greatly LF defense at Fenway. I think Fangraphs may degrade Bays defensive contributions too much.

[200tang]

These two comments are just spin city and the type of arrogance I was referring to. Others opinions are cast off because they are apparently "ignorant" in the minds of statheads, ie. "people who don't know what they're talking about". If its truly a blend of stats and scouting than one would have to agree IMO that the stats are flawed because no scout IMO will say that the 09 Cameron was worth more than the 09 Bay.

Good job making up info based on lies. Keith Law and Jason Churchill are two that come to mind, though.

[dickay]

here is some interesting stuff on Bay and fangraphs defensive reviews. Even god himself, by that i mean Bill James, says the results are inaccurate...or some of the posts i've found have stated. I have yet to find his quote on it. I"m assuming its out there since i've seen it now in multiple spots.

http://seattlesportsinsider.com/news/weekender

Fenway & LF
Submitted by Sandy on Mon, 12/07/2009 - 5:16am.

Personally, I put zero, zilch, nada weight on individual defensive stats coming out of Boston's LF. Some historical perspective. Looking at only the range portion of UZR.

Manny:


* 2002 - (1.8)

* 2003 - +3.7

* 2004 - (0.8)

* 2005 - (19.8)

* 2006 - (20.6)

* 2007 - (21.0)

* 2008 - (5.7) - combined from Boston/Dodgers

* 2009 - (2.9)


According to UZR, between 2004 and 2005, Manny lost 20 runs of range, which he maintained right up until the instant he was traded, at which point, he got back 17 of those runs.

As a comparison, let's look at his RF/9 over the same span.


* 2002 - 1.7

* 2003 - 1.8

* 2004 - 1.7

* 2005 - 1.9

* 2006 - 1.6

* 2007 - 1.7

* 2008 - 1.8

* 2009 - 1.6


It just so happens, that 2005, the year Manny morphed into the worst fielder in baseball, he set his CAREER HIGH in actual putouts, (243). UZR gave him a 7.0 ARM rating that year, (he had 17 assists, so not shocking).

Jason Bay was a plus fielder until 2007, (by UZR), then plunged toward Manny-land, with -14.6 and -14.4 range ratings since arriving in Fenway. Note that while Manny was posting consistent 1.7ish numbers, Bay posted a 1.8 in his first partial season in Boston, but posted a 2.3 RF/9 in 2009, and made 310 putouts, (and added 15 assists). Somehow, UZR says Bay COST the team 0.7 runs with his arm.

How many LFs in all of baseball posted a 2.3 RF/9 in LF? Three -- Crawford, DeJesus and Bay. To believe that Bay is *HORRID* in LF, I have to believe that a RF VASTLY above that of any LF in baseball would be required for him to be AVERAGE. Bay records the 2nd most outs of any LF in baseball, (trailing only Crawford), and is pegged as a bumbling bafoon by UZR.

Sorry, but if UZR says Manny was an average LF in 2004 with 198 POs, a 1.7 RF/9 ... and then contends Bay, with 310 POs, a 2.3 RF/9 in 2009 cost them more than 14 runs due to his poor range ... well, then UZR is a complete *****. (Or is anyone who chooses to believe UZR in this case).

According to Fangraphs, Bay was "expected" to make 327 outs. That would make him (in theory), a perfectly average LF. That was the most outs made by ANY LF in all of baseball. I am supposed to believe that Boston has what ... an order of magnitude more chances for the LF than any other LF in all of baseball? With the GREEN MONSTER?!?!? In what Bizarro world does LF in Boston become the place to pad DEFENSIVE stats? You have to go back to 1986 with Jim Rice to find a LF who had more POs for Boston than Bay.

Manny, at age 38, improves his range runs by 18 by leaving Boston. I'm thinking Bay -- going to Seattle -- probably ups his range runs by about 60. (No, that's not hyperbole ... that's my opinion on how reliable UZR range factors happen to be).

this was a follow up post in the thread;

But Bay, in 2009, was second only to Crawford in putouts, had a basically identical RF/9, and they played the same innings, (3 apart). I CANNOT accept that the 17 extra POs Crawford managed in 2009 explain the 32 run differential in range rating from UZR. It makes no logical sense.

[haveacigar]

These two comments are just spin city and the type of arrogance I was referring to. Others opinions are cast off because they are apparently "ignorant" in the minds of statheads, ie. "people who don't know what they're talking about". If its truly a blend of stats and scouting than one would have to agree IMO that the stats are flawed because no scout IMO will say that the 09 Cameron was worth more than the 09 Bay.

Ok, just because you think the other side has legitimacy doesn't mean you have to agree with them. C'mon now. I can accept a scout's opinion and then say based on stats that I disagree with them, and the disagreement doesn't make stats flawed anymore than it makes scouting flawed. In any case, judging someone else's opinion false isn't arrogant.

Anyway, scouts don't deal in "worth." That is a question that is almost only approached by statisticians. A scout would say "Jason Bay has much better power and contact hitting ability, while Mike Cameron has much better speed and fielding range ability" and let someone else make the "worth" judgment. The only people interested in the "worth" judgment are general managers and people who analyze roster moves. Scouts generally aren't analyzing roster management so much as the tools of individual players. On the other hand, most statistical applications of baseball involve analyzing roster moves, so that sort of analysis is generally exclusive to statisticians, management, and certain media members.

I don't know how you could have any discussion of worth in a given season without using stats anyway. How else are you supposed to make that determination?