Transaction Lab: Ryan Howard

by PWHjort on Apr 27, 2010 6:00 AM EDT in Trades & Transactions

Among saberists, the opinion of Ryan Howard's new 5 year, $125 million contract extension seems to be fairly universal. That being, basically, he's not good enough to justify the money, he's too old to justify the years, and his body, swing, and playing style are all likely to age poorly. Exactly what does this mean, though? Or, how badly did the Phillies get burned?

I used the slash line ZiPS projections Dan Szymborski published here in conjunction with the regression equation Adam Dorhauer published here to estimate Ryan Howard's weighted on base average going forward. With this, we've got something we can easily convert into run values, given his playing time. You know the drill from here. Add in fielding (bonus: use the UZR aging curve Jeff Zimmerman published here), positional adjustment, and replacement, and divide by the league's run-to-win conversion factor. And we've got Howard's projected win values--the basic commodity of baseball economics.

For the sake of generosity, let's assume Ryan Howard stays extraordinarily healthy for the life of his contract and accumulates 700 PA's, something he's actually done thrice, each year through 2017. By being generous with the playing time projections, we're being inherently optimistic about Howard's overall value. That's OK, what we're looking for is a sort of "best case" scenario. Here are the results:

And so with these projections in mind, how does the deal look for the Phillies?

It's easy to see why the sabermetric community has reacted so negatively. If ZiPS is even close to right--even if Howard stays unprecedentedly healthy--this deal isn't going to look good. Of course, nothing is out of the question, and we don't know this deal will end badly. But right now, that's where the smart money is.

Tweet Comment 11 comments | 0 recs |

Do you like this story?

Comments

$ per win

What coefficient are you using for X?

Val(M) = X * Wins

Looks like you used X = 4.2 for 2010 and it looks like your X is shrinking for each of the following years. I thought it was suppose to increase 10% per year? And if that is the case, you come up with a totally different conclusion. So I am just wondering what your methodology is here. Thanks.
vr, Xeifrank

by Xeifrank on Apr 27, 2010 3:30 PM EDT reply actions

It depends on the team's position on the win curve.

Regressed to 81 over a 5-year period. In Howard’s case, it was $4.1 in 2010, 2011, and 2012. $4.49 in 2013, $4.23 in 2014, and $4.1 in every subsequent year.

My calculator is adjusted for many things. You can read more about the Beta version of the calculator here. There have been some tweaks since.

Beyond the Box Score / Capitol Avenue Club / shwitter: @CapitolAvenue

by PWHjort on Apr 27, 2010 6:56 PM EDT up reply actions

I am just use to the method that Tango uses to look at these types of signings (no win curves). Using your projected WAR values and starting with a $/WAR value of $4.2M in 2010 and upping it by 10% each year I get a total contract value of $104.2M. Still shy of the $125M but a little more reasonable. Perhaps they have evaluated him higher than you have – let’s atleast hope so. :)
vr, Xei

by Xeifrank on Apr 29, 2010 12:58 PM EDT up reply actions

My calculator does the accounting for you, in a different way. Tango increases the price of a win, I adjust his salary to account for the time value of money—it’s the exact same thing.

I don’t think you can assume a 10 per cent rate of return in this economy, though. I used 8 per cent.

Beyond the Box Score / Capitol Avenue Club / shwitter: @CapitolAvenue

by PWHjort on Apr 29, 2010 3:29 PM EDT up reply actions

Thanks for taking the time to reply. I guess I need to study up on what it is you are doing. I just look at your chart and see the $/WIN dropping instead of rising and it doesn’t make sense to me. Let me show my work and please point out to me where I am going wrong. Let’s assume 10% inflation for the purpose of this example. I will round to nearest tenth.

Year, Wins, $M/Wins, Val(M)
2010, 4.0, 4.2, 16.8
2011, 3.3, 4.6, 15.2
2012, 3.2, 5.1, 16.3
2013, 3.0, 5.6, 16.8
2014, 2.6, 6.1, 16.0
2015, 2.1, 6.8, 14.2
2016, 1.2, 7.4, 8.9

Summing up the Val(M) column = $104.2M

So, if you and Tango are doing the same thing. Tango increasing the price of a win and you accounting for the time value of money – why the big descrepancy in results?

Even if I adjust the 10% inflation to 8%, I get $99.1M. To get a value of $78.6M, I have to adjust the inflation to negative 1.5%.

by Xeifrank on Apr 29, 2010 6:23 PM EDT up reply actions

OK, here’s how my calculator works. In addition to the Wins and dollars, I need three parameters. One is a team’s true talent level. Two is the rate of return on dollars (I generally use 8 per cent). Three is the rate of return on wins. I’m using 6 per cent for reasons explained here and here.

That last thing is probably the biggest difference between my calculation and yours. I’m using a discount rate x/((1+r)^n) WRT both the dollars and the wins. The best explanation I can give for why I came to the conclusion this is the correct thing to do is there’s a possibility that any team falls into an area where the marginal value of the win is nearly zero. Therefore, a win today (when you know what your position on the win curve is) is more valuable than a win tomorrow.

The other thing my calculator does is adjusts the price of a win for a team’s position on the win curve, such that the marginal value of the wins is greatest on the playoff bubble (83-88 wins, true talent level). This is regressed to the mean over a 5-year period.

You can download the beta version of the calculator here. Note I’ve used a base line of $3.5 M/W, so probably add about $0.6-$0.7 M/W if you’re trying to get what I did.

Beyond the Box Score / Capitol Avenue Club / shwitter: @CapitolAvenue

by PWHjort on Apr 30, 2010 4:04 PM EDT up reply actions

So lemme get this straight...

- Even if he stays completely healthy throughout the entire contract…
- Even if his body ages like an average player does…
- Even if he isn’t released before the end of this and they get $5m value from the picks when he walks…

…they’re still overpaying by $49 million. Classic.

On Twitter: @baseballtwit

by adarowski on Apr 28, 2010 8:13 AM EDT reply actions

I don't think #1 is an assumption here.

Not staying perfectly healthy is probably part of the “aging like an average player” piece. So if Howard stays perfectly healthy, his WAR per season will be higher than listed.

Beyond the Boxscore Not a member? Sign up.

by Sky Kalkman on Apr 28, 2010 9:37 AM EDT up reply actions

Adam is right. I gave him credit for 700! PA’s per year for the life of his contract.

Assumption #2 is the one I’m unsure of. I used ZiPS projections. I don’t think everyone ages the same under their projection system, I don’t know though.

Beyond the Box Score / Capitol Avenue Club / shwitter: @CapitolAvenue

by PWHjort on Apr 28, 2010 9:40 PM EDT up reply actions

Different types of players age differently.

Body type is a problem – I remain unconvinced that the crappy height/weight figures we have really gives us what we want (You end up with similar-bodied players being a whole lot of strange combos like Orlando Hudson/John Kruk and Cal Ripken/Charley Kerfeld).

--
Dan Szymborski
BTF
Dan on ESPN Insider

by D.Szymborski on Apr 29, 2010 10:17 AM EDT up reply actions