What's a Fair Return in a Roy Halladay Trade?
by Sky Kalkman on Jul 16, 2009 2:00 PM EDT
in Trades & Transactions
Email
PrintI have a great idea for a high-impact trade target that nobody's talking about: Roy Halladay. There would probably be a team or two interested, don't you think? If I see any further mention of this possibility, I'll let you know.
Pulling my tongue out of my check, dissecting the Roy Halladay rumors is a pretty fun exercise. While the AL East is stacked, it's not a foregone conclusion that the Jays are out of the running this year (or next), so any return offer would need to blow the Jays away with top prospects or include some quality MLB-ready talent. How do we weigh the value of Halladay versus a return package? Well, how about objectively, with a nifty trade value calculator?
Sure, there are contextual factors related to the Jays and potential trade partners to consider, but a context-neutral analysis is a great place to start. Let's take a look at Halladay's value and the potential value of offers from interested teams.
Roy Halladay's Trade Value
Halladay's under contract through 2010, owed about $7M more this year and $15.75M next year. A rough estimate of production (200 IP at a 3.00 ERA in a 4.40 ERA league) puts him at 6 WAR in 2010, and 3 more WAR in 2009. At $4.5M per marginal win, that production is worth $41M on the free agent market. Add in an expected $5M from Type A free agent compensation picks and a team would get $46M in production at a cost of $23M, or a net trade value of $23M. Not bad. I certainly couldn't afford that.
What's $23M Look Like in Prospects?
The Blue Jays are looking to acquire young players with more cost-controlled years than Halladay's 1.5. Whether the young players are minor league prospects or early-career major leaguers, the method of cost-savings is obvious -- these guys won't make any money their first few years in the big leagues and will be paid less then free agents during arbitration. Therefore, they don't have to be stars in order to help their teams as much as free agent stars, because they earn much less money. The Jays are trading value now for value later, when it will hopefully do more to help them reach the playoffs.
Thanks to Victor Wang, we have a good guess as to the future net value provided by prospects of various standing. Erik Manning compiled this chart:
| Top 10 hitting prospects | $36.5M |
| Top 11-25 hitters | $25.1 |
| Top 26-50 hitters | $23.4 |
| Top 51-75 hitters | $14.2 |
| Top 76-100 hitters | $12.5 |
| Top 10 pitching prospects | $15.2 |
| Top 11-25 pitchers | $15.9 |
| Top 26-50 pitchers | $15.9 |
| Top 51-75 pitchers | $12.1 |
| Top 76-100 pitchers | $9.8 |
| Grade B pitchers (as graded by Sickels) |
$7.3 |
| Grade B hitters | $5.5 |
| Grade C pitchers 22 or younger | $2.1 |
| Grade C pitchers 23 or older | $1.5 |
| Grade C hitters 22 or younger | $0.7 |
| Grade C hitters 23 or older | $0.5 |
It's apparent that a stud position player prospect would be more than enough of a return for Halladay on his own. Looking at Baseball America's mid-season Top 25, that could be Jason Heyward, Mike Stanton (not that that Mike Stanton), Jesus Montero, Justin Smoak, Buster Posey or Carlos Santana (not that Carlos Santana) . I'm not a prospect guru, but there are probably some tiers even within those guys. And a lesser position player prospect, someone between the top ten and top fifty of all prospects, looks to be exactly an even trade. Once you start talking lesser hitting prospects or pitchers, however, you're talking at least two. And moving down into the B levels (pitchers or hitters), you're talking about half a team of prospects in return.
Here's a quick and dirty estimate of what prospect packages would equal Halladay's trade value. Based on past trading history, some are more likely than others.
- One stud position player. (Win for Blue Jays.)
- One top 50 position player.
- Two top 100 position players.
- One top 10 pitching prospect and a grade B prospect.
- Two top 75 pitching prospects.
- Four grade B prospects.
- Twelve to twenty grade C prospects. (The Ricky Williams strategy?)
Addressing Specific Trade Rumors
Many teams appear seriously interested in Roy Halladay, including the Cardinals, Phillies, Angels, and Dodgers. Let's take a look at some of the fair packages each team could put together.
Cardinals
Brett Wallace (3B) appears to be the center piece of a deal, and as a top 25 position player prospect, would make for an even trade by himself. The Jays are asking for another pitching prospect or two, and while the Cardinals don't have a whole lot there after trading Chris Perez, it wouldn't take a whole lot more for the Jays to come out winners. Jess Todd (SP) and Daryl Jones (OF with good OBP/defense) are the only other prospect options at a B level or higher. I know Erik would laugh at the idea of Jones included with Wallace, though.
Phillies
While everybody's raving about the Phillies' pitching prospects, they only had one ranked in Baseball America's Top 100 before the season, Carlos Carrasco. Mid-season, however, Kyle Drabek has cracked the Top 25. Both of those guys would be woth $15M in excess value going forward, while B-prospects J.A. Happ and Jason Knapp are each about half as valuable. Just looking at pitchers, then, one from each group would make for a fair trade, while two of the studs or three of the four would tip things in the Jays favor. How Pedro Martinez looks over the next two weeks will go a long way towards determining how much of their future the Phillies are looking to sacrifice.
The Phillies do have some quality position player prospects, though, too, namely Dominic Brown and Michael Taylor, two outfielders who both made BA's mid-season top 25, plus Jason Donald (INF) who was a Top 75 pre-season pick. That's an expected value of $25M each for the two otufielders and $15M for Donald. Maybe the Jays strategy is to push for multiple pitchers, then "settle" for just one plus a blue-chip outfield prospect?
Angels
Brandon Wood. If you call him a B level prospect like John Sickels has, that's only $7M in excess value. Bump him up to a Top 100 prospect then he carries more like a $13M trade value. If you're even higher on him and assume he can be merely a 2 WAR player for six seasons, he'd provide $38M in excess value:
Getting a little crazy, if Wood is a 4 WAR player, that would be $76M in excess value. As these examples makes clear, prospects who pan out provide an amazing amount of excess value, and the only reason we don't expect them to be that valuable as minors leaguers is the high rate at which they don't live up to expectations. There's a lot of gambling going on with prospects, but the larger a stable of them a team accumulates, the more viable a youth-centered strategy becomes. [End of Poz-style aside.]
Being conservative with expectations, the Jays would need more than Wood in return for Halladay, and the other Angel options in the minors are sparse. Jordan Walden (SP) cracked BA's Top 75 pre-season, while Hank Conger (CA), Sean Rodriguez (2B), and Sean O'Sullivan (SP) were B-level prospects.
Dodgers
If I were the Dodgers, I wouldn't look to acquire Roy Halladay this season, as they're virtually guaranteed a playoff birth. With the flukiness of the post-season, the high cost of a trade wouldn't result in a significantly higher probability of winning the World Series. If they're thinking forward to next year, why not see if he's available in the off-season at a lower cost or put their money into free agents?
Final Thoughts
The great thing about analyzing trades like this is I'm making it quite apparent how I've valued (via help from others) Roy Halladay, various prospects, and young major leaguers. If you disagree, even subjectively, you can present your argument in the same language I did. Think Jason Donald is a Top 10 prospect? Change his value accordingly. Think Doc's heading for a serious injury? Drop his expected WAR and re-run the numbers. Now, you'd have to convince me those opinions are valid, but at least I know how our opinions compare.
To end, I'd be remiss if I didn't mention again that not all teams are in the same situation, both for 2009 and beyond. The Jays are a long shot to make the playoffs this year (although they still have a non-zero chance) while the Cardinals will see their playoff chances increase dramatically with each additional win. To the Cardinals, $20M in excess value two years down the road might not be as useful as $10M in excess value now, making a seemingly lopsided trade worth it. How do we handle that piece of the puzzle? Hopefully I'll be back in a few days with the answer.
57 comments
|
2 recs |
Do you like this story?
Comments
Phillies
at the end of the day and I am running the Phillies, I am making this deal happen. I start off offering two of these four (Dominic Brown, Michael Taylor, Carlos Carrasco and Kyle Drabek). If needed, I would eventually give up 3 of the 4. Roy Halladay makes the Phillies legit WS contenders for the next two seasons. You can replace prospects. You can’t replace WS Banners and with Halladay the Phillies could make a legit run at 3 in a row.
I don't disagree. I would want to quantify how much Halladay helps their playoff chances in 2009 and 2010.
And compare that to how much the prospects would help make the playoffs in 2010 through 2016, or whenever. Getting there…
It’s not like the Phillies go from 0% chance of WS to 100% chance of WS by acquiring Halladay, though. My WAG is 1/9 × 50% versus 1/7 × 75%
Beyond the Boxscore Not a member? Sign up.
We have an article going up in about, oh, four minutes about playoff probabilities.
They probably aren’t any better than BPro, but one cool aspect is that they’ll be open source in a day or two. You can plug in changes in expected WAR the rest of the way and get back an expected change in playoff probability. I asked Nick, who worked on the probabilities, to take a quick look at what adding Halladay’s WAR would mean to some teams. Cardinals playoff probability would go up about 22%! Braves? 19%.
Those are HUGE numbers and probably more than would be caused by having even a great prospect’s expected WAR and price savings (which would be spent on other value contributions) for a bunch of seasons down the road. Again, needs to be calculated.
Beyond the Boxscore Not a member? Sign up.
Looking at Baseball America’s mid-season Top 25, that could be Jason Heyward, Mike Stanton (not that that Mike Stanton), Jesus Montero, Justin Smoak, Buster Posey or Carlos Santana (not that Carlos Santana) . I’m not a prospect guru, but there are probably some tiers even within those guys.
I think trying to assign tiers to that crop is probably a case where you’d slice the rankings so fine that any distinctions made would be more or less meaningless when confronted with the uncertainty that prospects this young represent.
One thing that is going to help the Jays is that Doc is saying just the right thing at every turn to help JP maximize his return. someone suggests that Halladay doesn’t want to go to the NL to hit? Doc comes right back with a quote about tough AL lineups. He’s not demanding a super expensive extenstion to waive his NTC, and he’s only on record as really liking Toronto, but understanding the business side of things.
There’s also the idea that Toronto would deal within the division which is left out of the main article. Does the Trade Calculator have an opinion on dealing him someplace where they’d see him several times this season and next?
"I'd like to f*ck Sandra Bullock." - Pedro Martinez, explaining his secret ambition to Sports Illustrated for Kids.
No, the trade calculator doesn't take into account much context.
It’s a raw expected production vs. salary calculation. Working on adding some new pieces, but much of the context can be added “by hand” in a subjective, yet logical manor.
Beyond the Boxscore Not a member? Sign up.
im kinda surprise by how intuitive the trade calculator's findings are
those example packages look just about right in line with what will likely happen. funny, i was half-expecting to see some old-school sabermetric market imbalances exposed.
by Charlie Scrabbles on Jul 16, 2009 4:01 PM EDT reply actions
Well, the one spot I noticed inconsistency is with one really good (top 50, not 10) hitting prospect.
That should be fair right there, but isn’t enough to get something done in real life. It would take at least two B-level prospects, if not another Top 100 hitter or pitcher, which puts the numbers squarely in favor of the Jays side of the deal.
But yes, teams balk at a top 10 hitting prospect (which is correct here), and the other suggestions are pretty much in line, although most teams want at least one hot prospect instead of accepting four B-level guys.
Beyond the Boxscore Not a member? Sign up.
Value by time
Is there some adjustment for the fact that Halladay gives you immediate value while the prospects will impact in years i.e. there will be other factors which are now uncertain?
Also
The addition of lets say 3wins (or whatever amount) is much more valuable for a competing team like Phillies, Red Sox or Yankees, rather than Nationals, Padres or even Dodgers when they are so far ahead. So Halladay’s wins could be more important since the wins the prospects contribute could be to a non-competing team
This point is valid and I'm looking into updating the model to include it.
Nick’s playoff probability work will help in that.
As for future vs. present value, it’s addressed somewhat, but could probably be better. We’re using $4.5M per free agent win, which should go up going forward. But we’re also probably not downgrading future performance enough or at all. A 5 WAR player today is probably expected to be a 3 WAR player four years down the road. Conveniently, in the past, inflation and player regression have nearly canceled each other out. Winning now probably helps revenue more than winning later, however. Not sure how to quantify that. Although if we do end up converting everything into playoff probability (even in future seasons) then you could probably also convert that into revenue somehow. Nate Silver did some research on the value of winning more games and making the playoffs to revenue. Of course, the revenue curves will be different for each team (e.g. NYY vs. Pirates) so I’m not sure how fair it is to go solely by that, either. Maybe Victor Wang’s looked at this, he’s smart in these areas.
Beyond the Boxscore Not a member? Sign up.
What about playoff success?
Halliday may not be needed by the Dodgers to reach the playoffs, but he would be a crucial to being successful once there. In a 7-game playoff series, Halliday may get to pitch 3 times, depending on the schedule.
"What is that — five out of six? The numbers say what they say. They own us." -- Torii Hunter, 07/01/09
by NorCalRangersFan on Jul 16, 2009 4:59 PM EDT up reply actions
I addressed that in the article.
I’m sure he’d make a difference, but the playoffs are such a crapshoot that I doubt it’s as significant as most people think, especially given the cost. I could be wrong and it’s worth running the numbers.
Beyond the Boxscore Not a member? Sign up.
Aren't the playoff winners pretty much the best teams in the regular season?
Derosa.
by vivaelpujols on Jul 16, 2009 6:29 PM EDT up reply actions
Nah, playoff series area a near crapshoot.
Put a 95 win team against an 85 win team and the worse team will still win 40% of the time.
Beyond the Boxscore Not a member? Sign up.
I see a study here for playoffs. Chance to win depending of game won (or Run scored/Allowed) difference
Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.
by Jeff Zimmerman on Jul 16, 2009 6:52 PM EDT up reply actions
You can use the short version of the log5 (or the full version)
A 95 win team (Team A) has a .586 chance of winning a game against an average team. An 85 win team (Team B) has a .524 chance. (disclaimer: season winning percentage is assumed to be true talent chance to win)
The head to head chance of Team A winning a head-to-head game is .500 + .586 – .524 or .562. Obviously that means a .438 chance of Team B winning a game.
The chances of winning a 7 game series is a little more involved.
Let’s consider how Team B could win that series (I think):
Sweep: 1 way = .438^4 = .037
Win in 5 games: 4 ways = 4 × .438^4 × .562 = .082
Win in 6 games: 10 ways = 10 × .438^4 × .562^2 = .116
Win in 7 games: 20 ways = 20 × .438^4 × .562^3 = .131
Add them up to get the total: .366 chance of Team B winning a seven game series.
Methodology from here: http://www.baseballprospectus.com/article.php?articleid=4517
by Dan Turkenkopf on Jul 16, 2009 7:38 PM EDT up reply actions 1 recs
Which of course ignores home field and matchups (among other things), but it's at least a start.
by Dan Turkenkopf on Jul 16, 2009 7:50 PM EDT up reply actions
Home field is add .020 points of W% I think
Derosa.
by vivaelpujols on Jul 16, 2009 8:36 PM EDT up reply actions
Yes, but I would like to know if the Log5 formula still holds true in the playoffs
Derosa.
by vivaelpujols on Jul 16, 2009 8:36 PM EDT up reply actions
Why wouldn't it?
I can think of a few reasons:
Season winning percentage doesn’t capture true talent chance of winning – perhaps not, but you should be able to run the same calculation using whatever you want to calculate chance of winning the average game
Matchups affect the relative chance of winning – harder, but you could probably come up with a custom chance of winning each game based on the likely matchup
Momentum is important - Studes found that end of the season momentum doesn’t matter much, but maybe it does within a series. I think I’ve seen studies that suggest it doesn’t really, but I don’t remember where.
Did you have some others?
Because beyond those, I don’t know what would make a playoff head-to-head matchup any different from a regular season head-to-head matchup.
by Dan Turkenkopf on Jul 17, 2009 7:54 AM EDT up reply actions
World Series Records Depending on Regular season Record (last 100 World Series)
3 times the teams had the same winning %
49 times the team with better record won
48 times team with worse record won.
Will do division games in a bit.
Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.
by Jeff Zimmerman on Jul 16, 2009 7:22 PM EDT up reply actions
Didn't the Cardinals win
like 82 games in 2006?
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
My critique of this is that it fails to take into account "win leverage"
Just like games have situational leverage, so do seasons. A win is worth much, much more to the Giants than to the A’s right now, because the A’s benefit only from the more or less generic profit a team makes when it wins a regular season game (and conversely might actually hurt themselves slightly by getting a lower draft pick), and the Giants benefit from that profit AND from the increased chances that they reach the postseason.
I think you’re intuiting this by observing that the Dodgers don’t need Halladay (their win leverage is low because they’re far ahead) and by not talking about teams like the A’s (their win leverage is low because they’re far behind), but it needs to be taken into consideration in the numbers when calculating the value of having Roy Halladay for half a season.
If you really wanted to dig deep, you could also try to figure out win leverage for NEXT season, since even on April 1 the Pirates are (most seasons) going to have less win leverage than the Yankees will, but that might end up being impossible given free agent signings and departures, prospect promotions and so on.
Linda's in the cold ground, won't see her anymore
Somewhere out on the highway tonight, the drunken engines roar
It's just one of those things, one of those things
-- Al Stewart, "Accident on 3rd St."
In memory of Nick Adenhart and all victims of drunk driving
My last paragraph:
To end, I’d be remiss if I didn’t mention again that not all teams are in the same situation, both for 2009 and beyond. The Jays are a long shot to make the playoffs this year (although they still have a non-zero chance) while the Cardinals will see their playoff chances increase dramatically with each additional win. To the Cardinals, $20M in excess value two years down the road might not be as useful as $10M in excess value now, making a seemingly lopsided trade worth it. How do we handle that piece of the puzzle? Hopefully I’ll be back in a few days with the answer.
I addressed measuring everything in playoff probability in this comment above.
Beyond the Boxscore Not a member? Sign up.
Obviously my critique will disappear if you work through the numbers :-)
Linda's in the cold ground, won't see her anymore
Somewhere out on the highway tonight, the drunken engines roar
It's just one of those things, one of those things
-- Al Stewart, "Accident on 3rd St."
In memory of Nick Adenhart and all victims of drunk driving
Is anyone else wowed at how remotely low this is
I completely follow your analysis (and would agree with it and how you come down to the results), but wow.
Even when factoring in any possible impact on playoff probability and surplus value now v. surplus value later, still lower than what I would’ve expected in terms of sheer prospect haul.
><
Yes, I'm a bit surprised, but consider...
Halladay’s cheap for his talent, but not cheap: $15M a season. How often do those guys get traded? Tex wasn’t earning nearly that much a few years ago, was he?
Second, as we’ve mentioned many times, Halladay’s contribution (say 3 WAR) is much more important to the Cardinals or a playoff contender than the Jays (who are probably out of it). Think of it like leverage. Closers pitch when runs are the most important to winning. Well, teams who deal for Halladay need him when wins are most important to making the playoffs. So instead of $24M in value, he might be worth $40M in value. Oops, re-reading your post you mentioned that. Sorry. Leaving it though ;)
Beyond the Boxscore Not a member? Sign up.
by Sky Kalkman on Jul 16, 2009 10:18 PM EDT up reply actions
I think it winds up being low because the values that are set don’t count for a prospect not progressing as we’d like to see.
To wit, what would Felix Pie have been worth 5-6 years ago?
"I'd like to f*ck Sandra Bullock." - Pedro Martinez, explaining his secret ambition to Sports Illustrated for Kids.
You're having the same problem Will Carroll's having.
The numbers are based on past performance. They’ve accounted for the bust probability of prospects. A top 25 hitting prospect who lives up to his potential is worth, say, $100M in excess value. A bust, obviously, is worth $0 (well, less if you start from the beginning and include his likely signing bonus, which Victor Wang accounted for). You look at all the prospects to be ranked in a certain range and average all the career paths. That average yields an expected excess value of $25M for a top 25 hitting prospect. Compared to that $100M, $25M sounds pretty low, right? It’s that low because of how frequently prospects bust (or make the majors but don’t live up to expectations.) It’s just flat false to say that the values don’t account for a prospect not progressing like we’d like to see. Click through to Victor Wang’s research and read about it for yourself.
I and others, especially Erik, have been championing how flukey prospects (and draft picks to a larger extent) are on this site for a while. We know most people think too highly of prospects. That’s been tackled appropriately in the numbers.
You can’t prove your point with a one player example. There are tons of prospects who have busted and tons who have outperformed expectations.
Thanks for stopping by and thanks for commenting. I’m just a bit frustrated at the number of people recently who haven’t read the full article, the background research, or other comments and write things that are wrong or redundant.
Beyond the Boxscore Not a member? Sign up.
by Sky Kalkman on Jul 17, 2009 10:28 AM EDT up reply actions
Fair enough;
I made an off the cuff remark without thinking through the total analysis, I’m sorry if that is frowned on in this community; but that doesn’t change that the question Blicks asked is a fairly simple, yet deep criticism:
If your model is accurate, why does it not represent the package values demanded by a GM for a player like Halladay? Do you think there is any chance that JP would take Brett Wallace straight up for Doc, even though that would be a clear win?
"I'd like to f*ck Sandra Bullock." - Pedro Martinez, explaining his secret ambition to Sports Illustrated for Kids.
This is a good question:
If your model is accurate, why does it not represent the package values demanded by a GM for a player like Halladay?
Some of the answer is that Halladay’s production is extra important to a team in the race right now. $10M in production right now is probably going to be more valuable than $10M in production at some random season down the road. If you could choose when to score your 750 runs throughout the season, you’d win a ton of games. And since you can choose when to bring in your bullpen ace, his production becomes more valuable. Like I mentioned, this type of “leverage” analysis is important, and we’re getting to quantifying it soon.
Some of the answer is likely that some GMs wrongly project the value of prospects.
Some of the answer is probably that trading a guy like Halladay creates much more of a PR hit than acquiring prospects creates PR excitement. Even if bad PR doesn’t affect attendance like teams think it does, the fear of bad PR could increase the number of prospects needed in return.
Some of the answer is probably that value today is better than value down the road. That’s quantifiable, too, by someone with more econ-based knowledge than me.
=====
To answer your “off the cuff” question, throwing unproven hypotheses out there for discussion is certainly encouraged, as along as they’re presented as ideas to ponder and not fact. I was more annoyed that you were making a blanket statement about the research that was just plain false and wasn’t presented as an assumption or a question. If you click through to Victor Wang’s research, you can find his exact methodology. If I can off as too aggressive, I apologize.
Beyond the Boxscore Not a member? Sign up.
Valuation of variance?
The evaluation of the expected value of the prospects seems pretty impressive; however, in a player trade (much like in a financial trade), you are effectively trading both future value and risk for a less risky, current value commodity.
For Halladay, his WAR variance is relatively small compared to most pitchers. The WAR variance for prospects is very large. A “fair” trade would be E[Halladay’s Future WAR] = E[Prospects’ Future WAR]; however, the Blue Jays would implicitly lose that trade as there is additional value in risk, which is not captured by a first-order expected value.
Using an analysis like the above, would it make sense to review historical trades with the expected return on trade (as evaluated above) to attempt to determine what the associated “risk premium” is for trading a proven ML commodity for prospects?
Certainly doesn’t seem like it would be trivial in any sense with all of the factors associated, but sure does perk my interest.
Just a few thoughts from a baseball-crazed actuary..
Did you really mean this:
the Blue Jays would implicitly lose that trade as there is additional value in risk
The Jays would receive prospects, which are more risky, but you say they’d lose the trade?
I have more to ask you, but that’s important to clear up.
Beyond the Boxscore Not a member? Sign up.
Risk itself has value associated with it.
In the case of prospects and even Halladay himself, you’re talking about expectations of value with some variance associated with that value.
If Toronto recieved prospects exactly worth an expected value of $23M, that would be exactly Halladay’s expected benefit value; however, they would be trading a less variable $23M for a more variable $23M, and would be accepting additional risk at no cost.
Player A has WAR values of 6, -2, 2.
Player B has WAR values of 1.5, 2.5, 2.
Both players have an expected WAR of 2; however, Player A could have a really bad year, a really good year, or an average year. Player B is very consistently around the 2 WAR mark.
As a GM in a tight race, I’d want the consistency of 2 WAR over the variability of between -2 and 6 WAR.
None of this theory above takes into account the time-value of money/wins (I assume that is in Victor’s valuation of prospects), nor do I consider that a team that is further out in a race would benefit more from a lottery ticket than a team that is in the midst of a race.
Given that these types of trades often happen between teams that are on the opposite ends of the spectrum, it is conceivable that the premium associated with the risk nets out to nearly zero.
This is an avenue with great potential for study
I find it hard to believe that the actual risk premium is zero. Most of the “fair” trades according to these valuations probably wouldn’t be accepted by the Jays because they feel they’re not getting enough.
Some of that is likely improper valuation of the players in the deal, but some of it is the risk premium.
There might be a argument that the risk premium should be zero, but I’m skeptical it actually is.
I think it would be very hard to calculate the actual premium though. Not only do teams have very different risk profiles at any given time, but management will have different levels of aversion to risk. Combine those factors with the scarcity of these types of trades and it will be difficult to come up with a proper valuation.
Any ideas on how to approach it from an actuary? :-)
by Dan Turkenkopf on Jul 17, 2009 6:26 PM EDT up reply actions
on the risk front I think you have to have ask the question?
Would you trade for the following projected WAR/season:
2
2
2
2
2
or
0
0
0
0
11
If you were to make 5 trades you would be better off trading for people in the second group than the first group, since GMs can’t mess up or they lose their job they will alway go for people in the first group.
Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.
by Jeff Zimmerman on Jul 17, 2009 6:36 PM EDT up reply actions
or
the former is the kinda of guys you should trade for long term. while the later are guys you should gamble on short term.
Thinking not as an actuary but as an economist
Risk premiums for individual consumers are a form of revealed preference, which is to say they only become clear by analyzing trades where both participants feel they gained something material. Strictly speaking, it is the distance between the utility curve over the range of outcomes and the expected value curve (which is linear and is a function of the probabilities of various payoffs).
The rub is that the utility curve is what must be revealed, as you can’t go around asking GMs what their utility curves look like. But I suspect that Sky’s trade value calculator might show what the risk premium is.
For example, if we make the assumption that GMs are acting rationally in their team’s self-interest (several caveats there: GMs might be stupid, and they might act in their own, and not their team’s, self-interest), we could look at all the trades a GM has made, figure out how much he demanded above surplus in stud-for-prospects deals and how much he discounted prospects when trading them away, and infer an individual risk premium for each individual or perhaps each organization.
The only problem that occurs to me with this approach is that I’m not sure GMs make enough trades during their careers to even out the lumpiness of the data. But it’s a thought.
by Tommy Bennett on Jul 17, 2009 8:03 PM EDT up reply actions
Agreed
My degree is in economics (I figure that since Friedman was a failed actuary, if I don’t make it I’ll just turn to nobel prize winning economics work…) and agree with this.
It’s a good point you bring up that GMs would have differing risk profiles.
Due to that, a GM would need a significant number of trades to judge. And even then, their profile could change depending on the status of the franchise over that course of time — the win now vs build for the future approach.
This just keeps getting more difficult, blah.
Tough question
There are so many variables impacting the valuation of a trade, that narrowing it down to the risk premium would be pretty difficult.
I’d approach it by taking as many trades as I could get ahold of, using WAR evaluations for each side involved with some relatively simple assumptions surrounding the time at which the prospect would reach the big leagues in order to discount and inflate the value of the wins.
At that point, I think the remaining value transfer in prospects could be split between risk premium and marginal benefit of the wins transferred in relation to playoff odds.
To break that piece apart, I would have to know the additional expected revenue for making the playoffs, the odds of making the playoffs prior to acquisition, the odds of making the playoffs after the acquisition, and calculate the range of values assuming 0% of the playoff benefits are transferred in trade through 100% of the playoff benefits being transferred in trade.
Projecting the impact of prospects on future playoff revenue is nigh impossible, so the value of that would have to be assumed as negligible in trade talks, except for within intra-division trades (I would assume anyhow).
The risk premium would then be the only remaining value, effectively determining the risk premium by elimination – unless I missed something, which is entirely possible.
My previous answer was kind of round-about...
Short answer:
My thought is that if Toronto got exactly Halladay’s value back in prospects then it could be considered a loss since the variance in the prospects is higher than the variance in Halladay.
Or maybe Halladay's more of a 20 WAR guy...
The X factor is Roy Halladay, a pitcher the Rangers can afford both talent-wise and financially, and who would make them the favorites in the West over the Angels. Unless they make that kind of addition, however, they’ll slip under .500 by the end of the year.
Beyond the Boxscore Not a member? Sign up.
Yea it runs grit, heart and gut feelings.
and 2 year old scouting reports if you are Dayton Moore.
Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.
by Jeff Zimmerman on Jul 17, 2009 3:15 PM EDT up reply actions
Is Willie Bloomquist still 2?
I thought he just had his birthday.
Derosa.
by vivaelpujols on Jul 17, 2009 7:26 PM EDT up reply actions
Sorry it is a running joke that Dayton is using 2005-2006 number to evaluate talent
Guillen, Farnsworth, Freel and Betancourt.
Jeff Zimmerman - Protecting the world from RBI's and Wins from my mom's guest house.
by Jeff Zimmerman on Jul 18, 2009 12:49 AM EDT up reply actions
They still wouldn't be the favorites
Not even close
Derosa.
by vivaelpujols on Jul 17, 2009 7:25 PM EDT up reply actions
does this calculation account for the fact that the blue jays have their own farm system
and their own prospects?
thus, regarding the simplest trade possibility mentioned and therefore the easiest to analyze, halliday straight up for brett wallace…
wouldn’t one have to analyze the jays farm system and see what the chances are the jays would come up with a third baseman of wallace’s caliber within their own farm system and then
compare that to the chances of finding a pitcher like halliday in their own system
i would think third base is a relatively easy position to fill—- in that many minor league prospects can be taught the position relatively easily etc, notwithstanding the positions they start their careers in….therefore, arguendo, if the jays are set in every position except third base, and they identify four players on the farm who can be groomed to play third, then wallace’s value would have to be compared with the expected value of those four potential third basemen (as well as against average replacement player)
on the other hand, halliday is a pitcher…and possibly the best pitcher of his generation…there are five starters in a rotation and even the best rotation usually only features 3-4 top starters…assuming halliday is the best out there for this generation, there is virtually no chance that there is anyone in the cardinal’s farm system who can ‘replace’ him if the cards don’t sign him
as anyone who has somehow muddled through my confused thinking above can see, i am a bit unclear of what i’m saying here…
something about the finite lineup spots in a major league lineup and the ‘infinite’ amounts of possible replacements
I think there can be something to that.
If 3B was totally blocked, it might not make sense to trade for a 3B. But if two hot 3B eventually rise to the top of the farm system, one can probably be moved to 1B, 2B, or a corner OF spot. And if not, there’s always another trade.
Beyond the Boxscore Not a member? Sign up.
Maybe a concept of effective replacement?
That takes into account the likely replacement players from within the system? Most times when a team is looking to trade someone, they’re not going to replace them with a true replacement level player. Especially when dealing from a position of strength.
For example, although Halladay is a tremendous pitcher, because the Blue Jays have pretty good starting pitching depth, it might be less damaging to lose him than it would be for the Mets to lose Santana. (Not sure if this is true, but assume it is)
I have no idea how to measure this, or even if it makes sense.
As you can probably tell from different posts in this thread, I’m having trouble reconciling these trade calculations with the realities of trading.
by Dan Turkenkopf on Jul 21, 2009 9:16 PM EDT up reply actions
Effective replacement is a lot easier to calculate for the rest of 2009 than it is five years downt he road, obviously.
Matt Joyce is how RJ justifies trading Crawford for a pitching prospect right now.
Beyond the Boxscore Not a member? Sign up.
Yeah, but I think current effective replacement might be all that matters
Or basically, assume the current replacement will be the replacement for the rest of your prediction window, with a properly projected WAR.
That’s almost certainly not true, but otherwise you’re overwhelmed with unknowns.
by Dan Turkenkopf on Jul 22, 2009 7:53 AM EDT up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Beyond the Box Score
FanPosts
Community blog posts and discussion.
Recent FanPosts
Follow us on Twitter!
BTB: @BtBScore
Justin: @justinbopp
Satchel: @SatchelPrice
J-Doug: @RationalPastime
Daniel: @CamdenCrazies
Peter: @CapitolAvenue
Josh: @garik16
Julian: @splashinpumpkin
Matt: @devil_fingers
Bill: @BillPetti
JD: @JDSussman
Lucas: @DBITLefty
Jeff: @jeffwzimmerman
Lewis: @LewsOnFirst
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
