Dan’s Take: What Should the Capitals Do With Alex Semin?

What To Do With Alex Semin

Before I jump into the fray, there are a few things I wanted to note upfront:

(1) I don’t have any experience writing about hockey.

(2) And I’m coming at this from the perspective of a baseball analyst, so that’s where my base is. I know baseball and hockey are very different sports – especially since in hockey there is much more interaction between players and a team can be more (or less) than the sum of its parts.

(3) There will be cold, heartless calculations done. I don’t care a lick about the players (in this context). To me they’re just assets, like houses. You can have a really nice house – worth $10 M – but if the mortgage on it is for $15 M then it’s not actually very good for you. Hockey is a business, so there are revenues and costs (like players). If a player is worth $5 M but is paid $6 M, then that’s a bad deal for the team. If he’s paid $4 M, then the team has found $1 M in excess value.

(4) There will be numbers, many of which I make assumptions about and do hand-wavy things with. Precision would be great, but I don’t quite have the tools or the available data at my disposal to do things completely accurately.

Alrighty then. So, what should the Capitals do about Alexander Semin? I’m going to go through each option to the best of my ability, to see which branch bears the best fruit.

Option 1 – Keep Semin for the rest of the year and then let him walk.

How much is this worth to the Capitals?

Well we know that Semin is being paid $4.6 M this year and will be a restricted free agent after the season. I don’t know how much he’s worth as a player, but I can make an educated guess.

Based on this analysis, it looks like forwards make about $5 M per point per game (it’s for a pair of forwards, but a worthwhile starting point). Semin is at .86 pt/g career, 1.3 pt/g last year, and .94 p/g this year. Call it 1 pt/g, and you’ve got $5 M.

More substantially, using their VUKOTA projection system, Puck Prospectus had Semin as being worth 15.4 Goals Versus Threshold (goals above what a replacement level player would produce) in their pre-season projections – which was third on the team to Ovie (27.6) and Backstrom (17.4). With 6 goals being approximately equal to one win, and one win being worth about $2 M, Semin would be worth $5.13 M (how convenient). He was at 22.3 GVT last year though, which would be $7.43 M. If you split the difference to about $6 M, then that means Semin is “underpaid” (relative to his assumed free market value) by about $1.4 M this season. That is value that is accrued to the team. Prorating it going forward, and it’s about $1 M.

Then you need to account for – beyond his base production – how much he’d help the Capitals’ chances of getting to the playoffs this year and how much that’s worth.

Estimates for playoff revenue I’ve seen indicate about $1 M per game a team plays. Last year the average team playoff team played in a little over 10 games, so we could estimate that making the playoffs is worth about $10 M. Maybe tack on another $5 M for increased revenues due to a boost in attendance the following year(s), and an estimated average playoff bonus comes in at around $15 M. (This is about half of what it is in baseball, which makes some really loose intuitive sense as baseball teams have yearly revenues that are about double hockey teams’. Very loose sense.)

If we say that keeping Semin on the team increases the Capitals’ chances of making the playoffs (above what his replacement would do) by about 2% then he’s worth an additional $0.3 M to the team this year ($15 M times 2%).

The final piece to this puzzle is the compensation the Capitals would get when another team signed Semin. As a player with an assumed value of $5+ M, Washington should get back a first-round draft-pick, a second-round pick, and a third-round pick. To properly value the picks we need to know what kind of players they translate into. Based on this analysis, you’re getting – on average – an OK NHL player, a very good minor leaguer, and an OK minor leaguer (who maybe gets a shot). I don’t really know what that “means”, but I’ll assign values to them (production minus costs) of $1 M, $0.25 M, and $0 M respectively (hand waving is fun!). That’s total compensation of $1.25 M.

Add it all up, and the Capitals – who’ll be keeping Semin through the rest of the year and then letting him walk – are coming out ahead by about $2.55 M.

Option 2 – Keep Semin for the rest of the year and then sign him.

Most of the work for this one has already been down (everything through the compensation). Since Semin is a restricted free agent, one might expect the Capitals to be able to sign him cheaper than his free market value (maybe $6.25 M going forward). If they give him a 4 year, $24 M contract ($6 M a year) then they come out ahead by $1 M.

I have the excess value from signing Semin as about equal to the draft-pick compensation value for losing him, which would make sense if the system works properly.

There is some potential opportunity cost associated with signing him since the team might not be able to make other additions due to the salary cap, but I don’t consider that a big issue. The reason is, it doesn’t really matter where your value comes from (presuming you count it correctly). We’re trying to maximize team wins with a given amount of money, so if the same dollars would need to be spent on a free agent anyway it’s pretty equivalent (not taking into account what holes the team has). If Semin makes the team 2.5 wins better a season (costing $5 M), then losing Semin and using that $5 M to upgrade at 2 other spots by 1.25 wins each still leaves the team at +2.5 win. Not perfect of course, but the general idea shouldn’t be too far off.

So the total excess value to the team is once again $2.6 M. The difference between letting him walk and signing him depends largely on if he would accept a below market rate (and I think he might potentially be a little overpriced), and if that money could instead be used on even more undervalued assets and if the Capitals have places where they can improve easily (ie, big holes). I’d say they probably do, but what do I know?

Option 3 – Trade Semin at the trade deadline

Since Semin’s value is around $2.6 M, the team would need to receive more than that in a trade to make it worthwhile. They would take a slight hit to their playoff odds this year, but if they could get a couple quality players at positions of need who are still young and cheap, then I would say to go for it. If the general consensus opinion of Semin’s abilities is actually higher than his expected production, then you cash in that overvalued asset if you can.

Now if you assume that all NHL actors (as they say in economics) are rational and have perfect information and all that (ha!), then it would follow that Semin’s contract would exactly equal his value and what he would bring over in a trade would exactly equal to his value, and so the correct thing to do would be to keep him for the year and then let him walk (since that’s the only thing that would provide a surplus to the team).

In the real world, you keep him/trade him based on his specific contract demands and exactly what kind of package you could get back in a trade. I’ll leave that speculation as an exercise to the reader.

Daniel is a writer for the highly informative Fanball Sponsored Orioles Blog Camden Crazies, which you can view here. If you would like to comment on his statistics or offer any creative criticism, please comment below or email us here.

Tagged with:
  • As a SABR fan, I like where this is going, but I think with almost 1/2 the season played you have to look at what Semin has done and project what he will do the rest of the year vs. using VUKOTA. If we had a 2010 VUKOTA proj for Semin I think that would be more useful since we are paying for future stats, which in turn generates future wins.


    Agreed, the challenge with NHL stats is the lack of them in an easily digestible format.

    Nate Silver wrote a great article “Is Alex Rodriguez Overpaid” which could prob be ported to this argument, but again, we need current/future VUKOTA

  • Indeed neilg. I did the best I could with what I could find, but those stats don’t seem to be readily available anywhere.