Mets and Trades through time

[Guest AG/DC]

="Mark Healey"]Just speaking for myself, the math makes my head hurt. IMO, player evaluators are for the most part, usually judged on the amount of major league talent they are able to draft. For a GM with that tag (like Omar), i think the basis of judging the trade changes somewhat.

I'm not sure where you want to go. We're judging trades, not drafts.

="Mark Healey"]Let's say for instance, we look at the Cedeno/Dotel deal for Hampton/Bell. My first instinct is to give SP very little credit for this deal. A.) he failed to sign Hamton after acquiring him, which should have been the point in the first place.

Such a formula is designed to reflect that. Hampton's term of service ends with the Mets after 2000. Dotel and Cede�o continued to earn win shares within their term of service beyond that.

="Mark Healey"]That he broke down after signing with Colorado is a sympton of that terrible decision that Hampton made, not any insight of SP's.

The Colorado service isn't factored in, only 2000, just as the Jays get credit only for the time they controlled Cone.

="Mark Healey"]Derek Bell, after a solid first half, was a dreadful distraction in the second half that year.

I'm not sure what you're looking for here. We count his productivity against his salary, and accept whatever we humbly get.

="Mark Healey"]Conversely, the Mets dealt Dotel (who, arguable could have been their go-to releiver/closer in 2001, thereby perhaps sparing us the back-to-back Franco and Milk Baby blowups when the Mets were making their amazxing comeback that year.

And we'd count the impressive figures that Dotel put up on modest salary within the term of service the Mets dealt away right up until the end of 2005. Speculating on what might have happened is pointless. We're measuring the guys the Mets gained against the guys the Astros gained. I don't know what's with the Milk Baby thing and I'm guessing I don't want to.

="Mark Healey"]Plus, SP would have seen Cedeno first hand for the next two years, and would have realized that giving him a 4-year deal was laughable.

I don't get this either. He did see Cede�o. What are you hoping to factor in here?

[metsmarathon]

Mark Healey wrote: Conversely, the Mets dealt Dotel (who, arguable could have been their go-to releiver/closer in 2001, thereby perhaps sparing us the back-to-back Franco and Milk Baby blowups when the Mets were making their amazxing comeback that year.

i think you've got a long long way to go if you're going to try to frame any sort of argument wherein the mets, had they kept octavio dotel, and had the season unfolded in exactly the same way, would have pulled armando benitez from those braves games in favor of an unproven octavio dotel, who likely would never have had any save opportunities in the prior year, due in large part to the dominance of the actual closer. (and who posted a 5+ ERA in the real world with houston the prior year, netting only 16 saves)

if they did still have dotel, there's simply no rational argument to be made that he would have been handed the keys to the 2001 season over armando benitez, who, to that point, had helped carry the mets through august and most of september, closing big game after big game. but, really, that's a [silly] argument and a beaten-dead horse for another thread entirely.

[Guest Mark Healey]

I'm just saying that for me, that trades should be judged on the talent evaluation aspect, not terms of service or salaries.

[Guest AG/DC]

I don't really know what that means. How do you judge evaluation of ability except in subsequent display of ability?

[metsmarathon]

so, not production and cost of production, but flawed, anecdotal reckoning? go for it, i guess.

		year	2000	2001	2002	2003	2004	2005
player	total	a$	1.998	2.264	2.383	2.555	2.487	2.633
Octavio Dotel	582.31	W3^2(a/p)	65	225	224	35	32	1
	24.13	W3	2.8	5.5	6.4	4.7	6	1.2
	26.60	p$	0.24	0.305	0.435	1.6	2.8	4.75
Roger Cedeno	8.08	W3^2(a/p)	2	6
	2.84	W3	1.7	2.6
	4.30	p$	2.4	2.7
Mike Hampton	24.52	W3^2(a/p)	25
	4.95	W3	8.4
	8.40	p$	5.75
Derek Bell	1.44	W3^2(a/p)	1
	1.20	W3	1.9
	1.90	p$	5

dotel provided a terrific level of production-value, especially in '01 and '02. houston definitely came out as winners here.

keeping hampton wouldn't've helped even out the trade, as his production plummetted after the contract, while his cost skyrocketed.

[metsmarathon]

i'm starting to wonder if there's an inherent flaw, or perhaps if we're just uncovering an inehrent truth, that trading away cheap production for more expensive, albeit potentially greater, production doesn't seem like the way to "win" a trade...

[metsmarathon]

the frank viola trade (i'll post a chart later) comes out 832.3 to 36.1 against the mets. 23:1 against the mets

hampton went 23:1 against he mets as well.

and cone went 344:1 in favor of the mets.

this could be a fun, albeit tedious project! whee!

[Guest AG/DC]

Well, those extreme ratios seem to suggest you're using the wrong exponent. What is the figure next to "p$"?

[metsmarathon]

on the left, total warp3. on the right, yearly player salary in millions.

i think the problem is in the salary scale. if a player makes league average money ($3M for roundness) and produces 1 win above replacement, he would be considered a tenth as valuable as a player who produced 1 win above replacement, but for the league minimum ($300k for roundness), and eight times more valuable than a player who produces 1 win above replacement and makes the league maximum ($24M for roundness).

this means that production for production, a rookie produces has 80 times the value of an arod, thereby blowing up the scale to kingdom come.

this is driving the numbers to silly extremes moreso than the exponent used.

if i pull out the exponent, so that i'm looking at year-by-year value [warp3/(player$/avg$)], summed, i get 36.2:1 for the viola trade, 207.5:1 for the cone trade, and 31.61:1 for the hampton trade...

cubing warp3 gives me 16.2, 743, and 15.5, respectively.

and, finally, and i think i might like this better, as it drives down the dynamic range on the salaries, if i take the square root of the salary ratio, then sum the values, i get 12:1, 67:1, and 9:1, respectively. um, that means calculating for each year [WARP3 / SQRT (player$ / avg$)].

so now a league average player is about 3 times less valuable than a league minimum player per warp, and is also about 3 times less valuable than a league maximum player.

[Gwreck]

Are we sure about those Kent numbers? Cot's Contracts says he signed a 3 year deal with the Giants in 1999. Wouldn't that mean that the Mets only "controlled" him through the '98 season?

[Guest AG/DC]

Well, that's an absolutely valid point. We can measure these guys until the end of their current contract (or their reserve clause servitude for younger players), as you suggest, or through the end of the tenure that their current team is able to extend from that, as I've opted for.

I decided to go with the latter. The Mets signed Piazza with the hope of extending him --- and they almost didn't, as happened with Hampton. That they were able to extend one and not the the other feeds into the wisdom of the trade and therefore the subsequent evaluation.

Kent's contract with the Giants is an extension of his term of service that took him from the Blue Jays to the Mets to the Phils to the Giants. That he got extended by the Giants may not seem that relevant when we analyze the Cone trade, but it is, because

(1) if the Giants got him (at a relatively cheap price) to extend, it's specualtive to think he wouldn't have with the Mets or Indians (or, at least more speculative than concluding that he wouldn't have), and if we're going to give the Mets credit for win shares earned under Piazza's subsequent seven-year contract, we have to subtract win shares earned under extensions by players they traded away; and

(2) much more importantly, those win shares are going to be taken away from the Mets when we examine the trade that sent Kent and Vizcaino to the Indians for Baerga, so it would be unfair not to credit those shares to the Mets in judging Kent's acquisition.

[Benjamin Grimm]

I agree. Especially about Point 2. If the formula can be refined, AG's second point is where the magic kicks in.

[metsmarathon]

i decided to start with the 2007 trades and work my way backwards...

-	-	year	2007
player	total	a$	2.944556
Jose Castro	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	-
Sean Henry	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	-
Jeff Conine	(0.12)	W3*sqrt(a/p)	(0.12)
-	(0.10)	W3	(0.10)
-	-	p$	2.00

Jose Castro & Sean Henry for Jeff Conine
0 : -0.12 mets lose

drew butera	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	-
dustin martin	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	-
luis castillo	1.07	W3*sqrt(a/p)	1.07
-	1.50	W3	1.50
-	-	p$	5.75

drew butera & dustin martin for luis castillo
0 : 1.07 mets win

Brian Bannister	16.42	W3*sqrt(a/p)	16.42
-	5.90	W3	5.90
-	-	p$	0.38
Ambiorix Burgos	1.86	W3*sqrt(a/p)	1.86
-	0.70	W3	0.70
-	-	p$	0.42

brian bannister for ambiorix burgos
16.42 : 1.86 = 8.83 : 1 mets lose

Henry Owens	4.18	W3*sqrt(a/p)	4.18
-	1.50	W3	1.50
-	-	p$	0.38
Matt Lindstrom	8.35	W3*sqrt(a/p)	8.35
-	3.00	W3	3.00
-	-	p$	0.38
Jason Vargas	(1.11)	W3*sqrt(a/p)	(1.11)
-	(0.40)	W3	(0.40)
-	-	p$	0.38
Adam Bostick	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	-

Henry Owens & Matt Lindstrom for Jason Vargas & Adam Bostick
12.53 : -1.11 mets lose

Heath Bell	16.58	W3*sqrt(a/p)	16.58
-	6.00	W3	6.00
-	-	p$	0.39
Royce Ring	1.67	W3*sqrt(a/p)	1.67
-	0.60	W3	0.60
-	-	p$	0.38
Jon Adkins	-	W3*sqrt(a/p)	-
-	-	W3	-
-	-	p$	0.38
Ben Johnson	(0.28)	W3*sqrt(a/p)	(0.28)
-	(0.10)	W3	(0.10)
-	-	p$	0.38

heath bell & royce ring for Jon adkins & ben johnson
18.25 : -0.28 mets lose again!

the number next to the player's name is his sum value, while the number below that is his total production, in warp3.

to answer the question of who won the trade, you add up who the mets sent away, and if its higher than who the mets got, the mets lost the trade.

[Vic Sage]

metsmarathon wrote: i'm starting to wonder if there's an inherent flaw, or perhaps if we're just uncovering an inehrent truth, that trading away cheap production for more expensive, albeit potentially greater, production doesn't seem like the way to "win" a trade...

well, i think this is the inherent flaw in this line of thinking. GMs aren't trying to "win a trade"... they're trying to win championships.

Yes, the Mets "lost" the dotel/Hampton trade... but as a direct consequence of that trade, Hampton led the 2000 Mets pitching staff to the world series. So was it really a bad trade?

Depends on how you define "bad", i guess. But would the Mets have gone to the WS with Dotel in the rotation in 2000? He put up an ERA over 5 for Houston in 2000. How about more starts for Pat Mahomes? Would that have given us the 15 wins /200+IP /142 era+ that Hampton gave us that year?

And considering his performance since 2000, did the failure to re-sign Hampton for the mega-dollars he wanted (and got) really hurt the deal's value further, or does it help it? Yes, Dotel has had a few good major league seasons, but not that many. And considering Hampton was a frontline SP coming off a career year in Houston, I'd say we paid remarkably little for him, and were able to do so exactly BECAUSE he was unsigned beyond 2000.

The deal accomplished precisely what it was intended to. It got us to the post-season. The Yankees were simply better that year.

And any formula that considers the acquisition of Hampton in the winter of 1999 a "bad deal" needs to be reconsidered.

Furthermore, in establishing whether trades, in general, help more or hurt more, you need to look at each GM's record separately. Lumping Cashen in with Joe McDonald, with Jim McIlvaine, with Steve Phillips, together with Minaya is meaningless. they were operating in different environments, with different assets, and with different philosophies, and different owners.

I'd be interested in evaluating the moves made by Frank Cashen from 1980, when Doubleday/Wilpon bought the team until he left in 1990. The best team in Mets history was built with 1st round picks Gooden & Strawberry, and was impacted by the following trades (some bad, mostly good):

1981:
- Reardon > Ellis Valentine
1982:
- Greg Harris > George Foster
- Mike Scott > Danny Heep
- Mazzilli > Darling + Terrell
1983:
- Neil Allen + Ownby > Keith Hernandez
- Carlos diaz + Bailor > Sid Fernandez
1984:
- G.Young, M.Lee, M.Cook > Ray Knight
- Terrell > Hojo
- Hubie Brooks +Winningham,Youmans,Fitzgerald > G.Carter
1985:
- Jose Oquendo > Salazar, J.Young
- Schiraldi +Gardner,Christensen,Tarver > Ojeda
1986:
- Beane + Latham,Clink > Teufel

after that year, Cashen's deals included:

1986:
- Kevin Mitchell+abner,Jefferson > McReynolds
1987:
- Hearn+Anderson, Gozzo > Cone
- Orosco > Tapani, Whitehurst
1989:
- Aguilera +Tapani, West,Drummond > Viola
- Dykstra + McDowell > Juan Samuel
- Samuel > A.Pena + M.Marshall

[metsmarathon]

fun ruiner.

[nymr83]

]The Yankees were simply better that year.

i say "luckier", starting with the ball that cleared the wall that wasn't a homerun and payton(?) forgetting how to run the bases

other than that i agree with your post

[Guest John Cougar Lunchbucket]

Timo.

[Guest AG/DC]

1) I fully acknowledge going in that we have an additional challenge going forward of being able to weigh some win shares --- those that put a team into the playoffs, those in the post season --- as greater.

2) I never claim that the Mets inability to sign Hampton beyond 2000 hurts the deal. I do underscore the reality of the deal at the time it was made. Three things happen:

a) they resign him at an unGodly amount that makes the good he can bring in come at a dear price
they resign him at an unGodly amount and his career goes in the tank exactly as it did
c) confronted by these first two possibilities, they fail to resign him.

Confronted by these three, the deal has its limitations going at the time it was made. The Mets had to get their returns fast. Now we have to weight the returns.

The flaws in being unable to weigh sufficiently different types win shares is challenging. I think first we have to nail down the formula to compare under the initial assumption that all win shares are equal. You've got to start somewhere.

This is a flaw with win shares in general, in my opininion. Bill Mazeroski's homer against the Yankees is worth what? Fifty win shares? But it's not in his career total.

[Guest Rockin' Doc]

Payton?

Forget it, he's rolling.

[Guest AG/DC]

A quick guess says.

1981:
- Reardon > Ellis Valentine

Mets lose.

1982:
- Greg Harris > George Foster

This is actually Alex Trevi�o, Greg Harris and Jim Kern for Foster.

This is one that I guess the Mets win on win shares but the money spent on him per win share throws this back to Cincy.

Mike Scott > Danny Heep

Congratulations Houston. And congratulations to Met fans for not treating Danny Heep like crap over this.

Mazzilli > Darling + Terrell

Mets stomp Texas.

Neil Allen + Ownby > Keith Hernandez

Mets stomp Whitey Herzog.

Carlos diaz + Bailor > Sid Fernandez

Mets. What was Los Angeles thinking?

And where did Diaz disappear to after 1986. He was alternating good year/bad year and then the record stops after 1986.

- G.Young, M.Lee, M.Cook > Ray Knight

This is definitely one that the Mets kind of lose on raw win share numbers, as Young and Lee had the kind of long modest careers that can add up, but the Mets clearly won on Knight's 1986 post-season.


- Terrell > Hojo

Mets.

- Hubie Brooks +Winningham,Youmans,Fitzgerald > G.Carter

This deserves a close look. I bet it's pretty damn close. I don't believe for a minute the Mets can't win in 1986 with Hubie at short, but that's not what happened.

When Carter signed, he briefly became baseball's highest paid player.

- Jose Oquendo > Salazar, J.Young

St. Louis wins here. We gave up on Oquendo too young. Too bad Whitey rarely had a place to play him.

- Schiraldi +Gardner,Christensen,Tarver > Ojeda

I'll say Mets on 1986 win shares for Ojeda and loss shares for Schiraldi.

- Beane + Latham,Clink > Teufel

Bam! Mets.

- Kevin Mitchell+abner,Jefferson > McReynolds

San Diego wins. The salary helps.

- Hearn+Anderson, Gozzo > Cone

Mets smash hit.

- Orosco > Tapani, Whitehurst

Terrible deal for the Mets.

- Aguilera +Tapani, West,Drummond > Viola

Mets probably lose.

- Dykstra + McDowell > Juan Samuel

Mets are dumb.

- Samuel > A.Pena + M.Marshall

Mets are desperate. Probably still lose.

[Guest metsguyinmichigan]

Rockin' Doc wrote: Payton? Forget it, he's rolling.

Sweet Animal House tribute!

[Guest AG/DC]

Not that I've taken the time to understand the marathon forumula, but it seems an exponent is needed here.

Assuming you trade six guys for one. The money is even --- the one guy makes as much as the six combined --- if the one guy gives you 9 win shares above replacement and the six guys each give you 1.5, it would appear to be a wash. But it's not. The six guys are giving you your nine WARP pints but taking up six roster slots --- and a much bigger proportion of your plate appearances/innings pitched.

Either each win share above replacement should be given an increasing exponent, --- such as 1.0 for the first, 1.2 for the seconed, 1.4 for the third --- or the WARP3 should be divided by appearances some how that's fair for hitters and pitchers.

[metsmarathon]

i'd suggest that the way to do that analysis would be to look at wins above average, not replacement.

also, the other needed piece is how to adequately calculate value. right now, we've got low value as a positive number between zero and one, and high value as a positive number between one and positive infinity.

low value should be negative. high value should be positive. average value should be zero.

i'm stepping back from the notion that players should be grouped into various salary pools. when you're building a team, you're not required to have any number of players from each pool, just 25 (40?) players total. how you mix those roster spots up between the player pools is up to the team, as is how much each player should be paid.

so i would suggest that the appropriate thing would be to look at player salary compared to average salary.

i would define value as
WAA(actual) - WAA(expected)
where
WAA(expected) = expected WAA for a player who makes n times the average league salary, and
n = salary(player) divided by salary(league average)

[metsmarathon]

oh, i don't like my earlier method, btw. it needs negative numbers. i don't not like this new one yet, but give me time...