[Sky] points out that the Jays trading Halladay does hamper their 2009-10 post-season hopes but enhances them in future seasons (and hampers the Yankees’). This is true, assuming the prospects all pan out as predicted, which doesn’t always happen. As Matt Swartz pointed out in one of his submissions for BP Idol:

 ^^ In total, 51% of first and second rounds picks make the majors. ^^

So, the Jays would essentially be flipping a coin as many as four or five times, depending on the amount of top prospects they get.

Matt further points out:

 ^^ Of the 2052 players in the study, 1041 of them made the majors. Of those, only 109 players were traded and then debuted with a different team than the one that had drafted them. Of that group of 109, only 19 accumulated a WARP3 of 10.0 in their careers. As it turns out, for all the fans who scream at GMs for trading away the farm system, rarely do the GMs trade away impact prospects. ^^

Historically, the Jays would be running against the numbers if they were to trade the Jays to a division rival for a wealth of prospects. Probabilistically, if they are going to trade within the division, they should demand Major League-ready talent or bust.

Well, it depends on what you mean by “if the prospects all pan out as predicted”. I know prospects are a gamble, that’s why we need something like Victor Wang’s chart that takes into account both bust-rate and potential. (See recent Halladay article.)

When I’m talking about a prospect’s value, I’m not just looking at his ceiling (say, 5 WAR for Brett Wallace). I’m also looking at the odds he lives up to that potential. At 5 WAR, he’d provide $100M in excess value, which, well, is much better than Halladay’s $23M. But if you take “all the Brett Wallaces” (i.e. all the position players ranked in the top 25 of all prospects) then you find a large percentage are only 2 WAR players and a large percentage never provide value. So we hedge our bets appropriately and give Wallace a future value of “only” $25M.

Now, we don’t know what Wallace is going to do and you can’t even count on him. But put together enough pieces in a farm system and your variance goes down. If you have four “Wallaces”, you’ll probably get one stud and one regular.

Sorry, back to our tweet-off. Prospects are a gamble no matter whether you acquire them intra-divisinally or inter-divisionally. But they have the same expected value. Trading within the division just magnifies the booms and the busts. Using expected value seems like a decent route to me. And heck, in the AL East, the Jays are probably better off with high-variance players anyway. 90 wins won’t cut it.

UPDATE #2: Sky responds.

  ^^ Prospects are a gamble no matter whether you acquire them intra-divisinally or inter-divisionally. But they have the same expected value. ^^

The difference is that when you trade an impact player like Halladay within the division, you can confidently write in permanent marker that he’s going to provide his new team with around 6 WAR. Sure, he could get injured but he’s been very durable since 2006. There are a lot of sticks that could get caught in the gears, so to speak.

With the prospects, as Sky points out, it is a gamble. Going back to the figures from Matt Swartz, cited above, we have the following probabilities with four first- and second-round prospects:

    * 12.5% chance of all making the Majors; 25% three of four make it; 50% two of four; and 75% one of four.
    * 17.4% chance the prospect will contribute a WARP-3 of at least 10.0 over their careers (even less if you assume that many of the prospects end up playing with different teams than the ones they were traded to).

(Note: WARP-3 and WAR are not interchangeable. That may be obvious but I feel it’s worth pointing out.)

So, if the Jays trade Halladay to the Yankees, they are almost definitely going back 6 WAR in the division race in at least the next two seasons. Halladay, at 32 years old, is going to start to decline at that point, so we can’t expect 6 WAR every season until he retires.

At any rate, the odds of the prospects providing an average of 3-5 WAR from 2011 until the end of Halladay’s career are not in favor of the Jays. To make an analogy to poker, the Jays making this trade to a division rival is like chasing an open-ended straight draw. If the Jays are going to gamble, they may as well gamble without spotting a division rival an average 6 WAR in 2009 (prorated) and ‘10, and anywhere from 3-5 in ‘11 and a couple years beyond.

There’s also the finances to consider: Halladay costs money, $15M next year, even to the Yankees. That’s $15M they can’t spend elsewhere, which could buy a bit over 3 wins. So, the edge to Yankees is only 3 WAR, not 6 WAR next year. And that’s only an edge for one season. (Plus this year, yes.)

But the Jays would get six years from a prospect or 4-5 years from a young major leaguer. Even at an advantage of only 2 WAR per season, that’s 8 to 10 total WAR over the contract, more than the Halladay edge.

And again, I understand that prospects flame out a lot. But we’re not talking draft picks here. Draft picks are young and mostly not MLB-ready. To rate highly as a prospect you have to be closer to the big leagues or have absolutely huge potential. The 30th ranked prospect right now is WAY better than the 30th pick in the draft. And the closer you get to the top of the list, the lower the flam-out rate. Top 10 hitting prospects are actually quite projectable. Top 10 pitcher prospects? Not so much.

But all that is accounted for in Victor Wang’s model: all the probabilities, all the potential levels of production, all the possible career trajectories. The production is compared to payment for each probability, then all are combined. When I quote a number saying a prospect is expected to return $25M in value, that’s not best case scenario. That’s a weighted average of best-case, worst-case, and every case in between.

On the “single year of Halladay scale” we’re talking about two different ways of expecting 6 WAR. One way, you know you’re getting it (Halladay). The other way (prospects) you might get 9 WAR and you might get 3 WAR. 6 WAR isn’t the ceiling, it’s the average. So, which would you rather have, a definite 6 WAR or a highly variable 6 WAR? Well, they’re pretty even, with maybe a small advantage to one or the other. And remember, while landing the 3 WAR is a double loss to the Yanks, landing the 9 WAR is a double win.

And while one prospect might be highly variable in expectation, in aggregate your risks go down. That’s why anyone can win one poker hand, but the pros win in the long run. The analogy that the Jays are chasing an open-ended straight draw is a poor one. Given the right return of prospects, this could be like playing AK suited against a pair of nines in Texas Holdem. It’s nearly a coin flip, but advantage to the AK even though the pair is the known commodity. In the long run, you’ll be happy betting the AK every time. And given that the Jays are the underdog, they should want to take chances.

Hi all, thanks to Crashburn for highlighting this issue and alerting me to this conversation. It’s right up my alley. I’m going to take Sky’s side on this but for different reasons I think than Sky is saying.

Firstly, Crashburn is absolutely right that trading Halladay to the Yankees basically doubles the cost in lost WAR for this year (not quite, because it may be that the Red Sox or Rays are the team to beat in those years, but you get the idea that it’s more valuable). However, that WAR is not worth the same dollar value to both teams. It’s worth more to the Yankees who stand a chance to compete in 2009 and 2010 unlike the Jays. Now, what is also true is that the value of the prospect doubles on the other side so even though that value comes with uncertainty (more on this in a second), it is still double the value to the Jays in subsequent years to not only get the benefit of those prospects but to take them away from the Yankees. So what this does is double the trade value— double the value for the Yankees now which they value relatively more than the Jays, and double the value for the Jays later which they value relatively more than the Yankees. The Jays need to damage the Yankees later on to have some sake of competing with them.

Try this as a thought experiment. Suppose that the Red Sox and Yankees arranged the following trade. The Red Sox gave the Yankees all their best players for nothing for 2009, and the Yankees then traded them all back with all of their best players in 2010, and they alternated like this year after year. Would the teams be better off? If you go by straight WAR, this is a neutral trade. But it’s not. Doing something like this would create huge juggernauts every year to go along with a bad team. This would probably create more combined value because you’re almost guaranteeing a title to one team each year. That has more value then both teams being borderline playoff teams. This might not make as much sense with the Yankees and Red Sox because they play in a division that frequently produces the AL Wild Card, but you can try with any two very good teams in a division and get the same basic structure and outcome.

Crashburn points out the randomness of prospects plays a role, and Sky counters that the expected value is the same. True, true. It works the other way though. The expected value is the same, and therefore the trade value in a neutral setting would be the same, but the variance of the outcome is higher. This is good for the Jays. Let’s say the Yankees trade a set of prospects to the Jays that yield about 5 WAR on average for 2011, but really its a 50-50 shot of 10 WAR and a 50-50 shot of 0 WAR. Let’s say the Jays were expected to win 80 +/- 5 games in 2011. Now they are suddenly expected to win 85 +/- 10 games. If the Yankees were expected to win 90 +/- 5 games, adding variance would help the Jays. They would rather add a 50-50 shot of 10 WAR than a 100% chance of 5 WAR. They want to increase the variance.

Go ahead and spin my argument back at me. The Yankees had some variance in their win total too with those prospects on their team, right? So all it does is lower the variance for the Yankees and increase it for the Jays making the difference between the Yankees and Jays have the same variance, right? Not quite, because the Yankees frequently trade these prospects as should good teams who are expected to win enough game to make the playoffs. Why add variance? Trade it out of the division. Maybe it’s the Yankees who should be afraid to trade for an ace within the division.

the only part of trading in division I don’t like is that some players play better when they put their mind to it. Like when you play against a team that tarded you away you play harder to get them back.

“well yes those prospects are worth a lot, but the might become busts and be worth nothing!” argument is so pervasive. These are probabilities we are dealing with, and the probabilities of boom or bust are already factored into the values.

Also, I wasn’t sold on intra-division trading at first, but you’ve definitely made an excellent point. While the Jays technically lose 6 WAR relative to the yankees, they lose 0% playoff probability, and after Halladay’s contract ends and on the Yankees lose 5-10 WAR even as the Jays gain it, resulting in a possible 20 WAR swing. Dear god.