Mets and Trades through time

[Guest AG/DC]

The first trade in Mets history was something of a steal. On Novmeber 28, 1961, They got Frank Thomas and a player to be named later for cash and a player to be named later. The player that completed the trade for the Mets was Frank's outfield-mate and procreative likemind Gus Bell, who went over on May 21, 1962. The Mets got back Rick Herrscher.

The problem is that I don't know the salaries of the players involved, nor how much money the Mets sent, but I still feel confident in calling it a Met win.

	Gus Bell	for	Frank Thomas	and	Rick Herrscher
Term of Service Ended	5/12/1964		4/5/1966		1962	Term of Service Ended
Win Shares	1.6		12.6		0	Win Shares
$	$?		$?		$?	$
Avg $	$?		$?		$?	Avg $

[nymr83]

what are your criteria? is it what the guys do for their new teams? what they do for the rest of their careers no matter where they play? these questions need answering.
does it matter if a guy contributes to a playoff team while the guys he was dealt for were still too young to help?

[Guest AG/DC]

My criterion is win shares through the term of service under which the player is dealt. If they continue to re-sign while under contract, they are eligible to continue piling up win shares towards the trade, even if having been traded again. If they are released or leave as free agents, the books close.

[Benjamin Grimm]

So you're saying that in the trade that sent David Cone to Toronto for Jeff Kent and Ryan Thompson, we'd only look at what Cone did for the Blue Jays and what Kent did for the Mets?

I can see that, but the problem with it is that I'd say that the Cone for Kent trade (let's ignore Thompson for now) was a very good one; they got a young player who ended up on a Cooperstown path. Should the fact that they later traded Kent in a bad deal reduce the quality of the original deal that imported him?

[Guest AG/DC]

Benjamin Grimm wrote: So you're saying that in the trade that sent David Cone to Toronto for Jeff Kent and Ryan Thompson, we'd only look at what Cone did for the Blue Jays and what Kent did for the Mets?

No, we'd look at what Kent did through the end of his term of service, which would take him through the 2002 season when the Indians lost control of him. Note that I wrote above, "even if having been traded again." And I tracked Frank Thomas through his release in 1966.

[Benjamin Grimm]

Ah! I see. That makes sense, then.

[Guest vtmet]

Benjamin Grimm wrote: So you're saying that in the trade that sent David Cone to Toronto for Jeff Kent and Ryan Thompson, we'd only look at what Cone did for the Blue Jays and what Kent did for the Mets? I can see that, but the problem with it is that I'd say that the Cone for Kent trade (let's ignore Thompson for now) was a very good one; they got a young player who ended up on a Cooperstown path. Should the fact that they later traded Kent in a bad deal reduce the quality of the original deal that imported him?

I don't know how I'd really score that trade...With the way that things went for Kent as a Met/Indian, I don't know if he would have turned out as good as Kent the Giant if he remained on the Mets, he just never fully seemed to fit in here...Kent/Everett had talent, but fit the adage that it's hard to be a young player in NY....And while Cone's Mets weren't exactly producing a good record at the time, he still was a very effective pitcher (7 complete games, 5 shutouts, 2.88 ERA, 214 strikeouts in under 200 innings in his 27 starts as a Met in '92 prior to the in-season trade)...I'd have to vote that we lost miserably in that trade, based on Cone's performance as a Met, Kent's performance as a Met, and Cone's performance post-Methood...

[Guest AG/DC]

Thanks, but it's just not working.

Let's dial back to 8/27/92.

	David Cone	for	Jeff Kent	and	Ryan Thompson
Term of Service Ended	8/30/92		12/18/2002		12/20/1996	Term of Service Ended
WARP3	2.6		69.9		12.2	WARP3
$ (in Millions)	$0.779168		$33.097483		$0.491983	$ (in Millions)
Avg $ (in Millions)	$0.198808		$15.946443		$3.578770	Avg $ (in Millions)
WARP3/($/Avg$)	0.66	for	33.68	and	88.74	WARP3/($/Avg$)

As expected, the idea that all win shares are equal hurts this measurement, as Cone provided win shares that perhaps helped put Toronto over the top. Are they then worth more? He then threw 22 solid innings for Toronto in the post-season.

More distorting (and distrubing) is the overvaluing of the relative value of modest prodution of a guy like Ryan Thompson at near big-league minimum salary. I wonder if we can craft such a thing as a replacement-level salary that we can compare them to instead of the average salary, since we're comparing them to a replacement-level player.

The Mets won this trade, but not that well.

[Guest AG/DC]

Regarding vtmet, I just think it's pointless specuating on what we feel would have happened. What was gained on one end and what was gained on the other are knowns that we're trying to measure.

[Benjamin Grimm]

Yeah, I think that if your formula shows that Ryan Thompson brought more value than Jeff Kent, then something is clearly wrong with the formula.

I'm not sure I understand the money figures. On the "$ (in Millions)" line, it looks to me like Kent made $33 million through the end of 2002, and Ryan Thompson made $491,000 through the end of 1996. If that's the case, I don't see how his average salary is $3.5 million.

[Guest AG/DC]

Those are the earnings of the average Major League player during that period.

[nymr83]

i just don't think it makes much sense to take money into account in this way with the Mets. clearly the Mets have tons of it, so just because A produced more wins per dollar than B doesn't make A for B a bad trade for the Mets, it might even be a good trade if B produced more wins even at a far higher cost. i'm not sure how i'd do it though.

[Guest metsguyinmichigan]

I weep at the thought of the Staub for Lolich trade.

The Seaver trade, while practically destroying my childhood, probably won't work out that poorly on paper, since Flynn and Henderson were everyday players and Zachary wasn't all that horrible.

Not saying it wasn't the darkest day of the franchise -- it was -- but I think did more damage morally than statistically.

[RealityChuck]

Nymr83 wrote: i just don't think it makes much sense to take money into account in this way with the Mets. clearly the Mets have tons of it, so just because A produced more wins per dollar than B doesn't make A for B a bad trade for the Mets, it might even be a good trade if B produced more wins even at a far higher cost. i'm not sure how i'd do it though.

I agree. Just stick with on-field performance. If someone helps you win the World Series, it doesn't matter if you paid him the league minimum or $20 million a year. Adding the money only complicates trying to evaluate the trade.

[Benjamin Grimm]

I can see that. But the problem is, a lot of times money is the underlying reason for the trade. So factoring in the money helps judge the overall success of the transaction.

But from a fan's standpoint, I am more interested in the impact on the field rather than the impact on the books. I can go either way on this.

The dollars are definitely the thing that's skewing the Ryan Thompson score.

[Guest AG/DC]

It's easy for us to say money doesn't matter. It does.

Blowing all of it on one guy keeps you from getting another better guy. Or lots of better guys.

It also forces them to keep these guys around longer than they would have.

David Cone was traded by the Mets for the sake of salary. The Jays didn't get anything more out of him due to salary.

[Benjamin Grimm]

I'm not sure how to correct the formula then.

Maybe the Ryan Thompson numbers are an aberration that makes a good formula look faulty? Should we try a few more transactions and see how they turn out?

Some possibilities would be the trade that brought Kevin McReynolds from the Padres, Steve Henderson for Dave Kingman, Jerry Koosman for Jesse Orosco, Matlack and Milner to Texas, Hebner for Espinosa, Reardon and Norman for Valentine, and Mazzilli for Darling and Terrell.

(I was trying to pick a variety of trades that all involve players whose careers have ended. I don't know how far back salary information goes, however.)

[RealityChuck]

AG/DC wrote: It's easy for us to say money doesn't matter. It does.

Really? If a guy gets the league minimum and and 20 win shares, and another guy gets $20 mill and 20 win shares, there's a difference in the team's performance?

]Blowing all of it on one guy keeps you from getting another better guy. Or lots of better guys.

But that's impossible to quantify. How do you know what better guys you could have gotten with the money? Maybe the money would have been spent on getting worse guys.

]It also forces them to keep these guys around longer than they would have.

Sometimes. Sometimes not. Besides, that's length of contract, not salary.

]David Cone was traded by the Mets for the sake of salary. The Jays didn't getting anything more out of him due to salary.

No. What they got out of him was wins and losses, which they would have gotten out of him no matter what the salary. If a player helps the team, he helps despite his salary; if he hurts the team, he hurts it no matter what the salary. Salary is a factor in trades since free agency, but so is "team chemistry," whether a player is a "good citizen," whether he show promise, etc. You might as well factor those in, too.

[Guest AG/DC]

In short.

1) Yes, there typically is.

2) No, it's not impossible to quantify. If money saved goes towards poor decisions, that counts against the evaluation of those poor decisions. We proceed on the assumption that money has an average value in wins, and good deals are when you exceed that. Bad deals are when you don't.

3) Salary without performance is a burden and a detriment to a team, over a year, over ten years.

4) I don't know what "What they got out of him was wins and losses, which they would have gotten out of him no matter what the salary" means regarding my position. What I tried to express was that they weren't able to get any more wins out of him due to salary demands which exceeded thier budget. He left the team as a free agent.

[metsmarathon]

i think the problem with the formula that i didn't recognize at first is that it is only measuring production versus cost of the players involved in the trades - essentially 'value'.

but when you're trading for a player, you don't just want value - you also want total production!

is there a solution? i dunno. maybe squaring the WARP3 would do the trick, thereby measuring total production X production rate. in this case, it might be more appropriate to do it on a year by year basis, and sum the results, as well.

kent would be (i'll assume $2M avg salary here for expediency):

{0.8 x [0.8 / (0.11/2)]} + {3.1 x [3.1 / (.19/2)]} + {5 x [5 / (.4/2)]} + ... = some number

as compared to simply:

69.9 x [69.9 / (33.1/15.9)] = 2,347 for kent

and

12.2 x [12.2 / (0.49/3.58)] = 1,087 for thompson

the latter method is sure easier, but i think the former might be more correct, as you're not squaring the sum, but rather summing the squares.

3^2 + 3^2 = 18
(3+3)^2 = 36

i'm overall not sure if that'll make things better, or just more cumbersome, but if you've got the yearly salary numbers (i don't, necessarily) and the time (i don't really) and the inclination (not me), it might be worth an investigation.

[A Boy Named Seo]

metsmarathon wrote: ...but if you've got the yearly salary numbers (i don't, necessarily) and the time (i don't really) and the inclination (not me), it might be worth an investigation.

Love it.

[metsmarathon]

perhaps another method, and i've seen snippets of this across the internet, would be to come up with a wins above contract type of figure.

basically, you would assume there to be three different types of players out there - pre-arbitration, mid arbitration, and post-arbitration - which have different wins/$ rates. the pre-arbitration list would be rookies, two year players, and three year players, right? so you figure, for a given year, what the total WARP3 for these players is, and their total salary, and divide to get a wins/$ rate for their salary. actually, to be more correct, you'd prolly do salary above league minimum (essentially replacement level) since you're looking at wins above replacement, and this might keep em on the same scale.

for the mid-arbitration players, you'd look to all the players who are in their arbitration-eligible years, and do the same thing - total wins above replacement divided by total salary above league minimum. then repeat again with all players beyond their 6th year of service - all those who are eligible for free agency.

these three populations get paid at markedly different rates, i'm sure you'd agree, making comparisons really only valid within the groups.

then you could take any player in any given year, and figure out which group he belongs to, and figure out, based on his salary, how many wins above replacement you would expect him to have. the difference between this and his salary would tell you the approximate value you're getting from the chap.

mind you, upon further reflection, i think this might be more of a measure of how effective and efficient your GM is as opposed to being able to measure the merits of one side of a trade against the other... but you tell me.

i do know that this would be virtually impossible, i think, to figure out without working directly for baseball prospectus, and having access to their entire database instead of typing in one-by-one all the player names to get the WARP3 figures... (is such a thing available, btw, or is the manual entry 'DT cards' the only way to get at the WARP3 data?)... so the other way i've described sure looks better right about now!

[Guest AG/DC]

Run with it. I obviously can't make my formula work.

[metsmarathon]

if you can tell me where to find a) minimum league salary by year and average league salary by year, i can run through my first method, which is merely an attempt at improvement of your formula.

the latter method is more a philosophical approach for somebody with far more time and access to attempt, and which may have already been approached by some out there, just not comprehensively, and not applied to the mets.

i can start with kent/thompson for an initial sanity check.

[Guest AG/DC]

Here's B. Thank you, CBS.

I'll guess that A is available at the MLBPA site somewhere.

[metsmarathon]

cool. the league minimum isn't as important to me... i suppose there's just no real reliable salary data available prior to '89?

[Guest AG/DC]

I'm sure there is somewhere, but that certainly gives you the years you need to test any approach on the Cone-for-Kent/Thompson trade.

[metsmarathon]

		year	1992	1993	1994	1995	1996	1997	1998	1999	2000	2001	2002
player	total	a$	1.084	1.120	1.189	1.071	1.177	1.384	1.441	1.720	1.998	2.264	2.383
David Cone	1.72	W3^2(a/p)	2	-	-	-	-	-	-	-	-	-	-
	1.31	W3	2.6
	2.60	p$	4.25
Jeff Kent	348.19	W3^2(a/p)	6	55	68	28	10	23	26	17	40	37	37
	18.66	W3	0.8	3.1	5	4.3	4.1	6.6	7.9	7.6	10.9	9.9	9.7
	69.90	p$	0.109	0.195	0.437	0.71	1.96	2.625	3.4	5.75	6	6	6
Ryan Thompson	244.53	W3^2(a/p)	8	62	135	40	-	-	-	-	-	-	-
	15.64	W3	0.9	3.1	5	3.2	0
	12.20	p$	0.109	0.175	0.22	0.275	0.275

so, here's what i show. for each player, i found their wins above replacement, their salary, and the average salary in the MLB at that time.

i calculated, for each year following hte mets trade until such time as the player was either released, left his team via free agency, or retired (a player who is granted free agency but returns to his previous team is not considered to have left his team), the players' production times the players value.

production is defined as wins above replacement player. value is defined as wins above replacement, divided by the ratio of the player's salary to the average league salary.

this turns into, for each year, WARP3^2 * average MLB salary / player's salary.

i've summed this for each player - the number immediately to the right of the player's name. the number below this is the square root of the above sum, which is an easier scale to look at, but is likely meaningless. the bottom number is the sum of the player's WARP3's, or total production.

you'll note that the bulk of thompson's production-value comes in 1994, where both he and kent produced 5 wins above replacement, but he did it at half the cost.

right now, this puts kent:thompson at 1.42:1 (an excursion i just did, with warp3 cubed, puts it at 2.17:1, but i have a harder time framing that logic.)

and maybe this does tend to drive home the point that part of a trade is getting or giving away cheap talent...

[Guest AG/DC]

I can't sort through the math right now, but that seems more like it.

[Guest 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.

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

Derek Bell, after a solid first half, was a dreadful distraction in the second half that year.

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.

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.

Oh well, maybe not. :-)