The price of a win in the 2016-2017 free agent market

Televised baseball is back! It feels like the winter is so far behind us now. The large majority of useful players have found teams, though Joe Blanton continues to look for work. At this point of the year, we’re mostly looking forward to the upcoming season. For a second though, we’ll look back at the off-season that just transpired. Our goal today is to find the price per win of the latest free agent market.

For all signings, I used the Free Agent Tracker at MLB Trade Rumors. They had a total of 101 free agent signings, including Jeremy Hellickson and Neal Walker’s acceptance of the Qualifying offer. I also utilized my average aging curves that I developed based on Fangraphs data from 2006 on. For instance, here’s the average aging curve for hitters, on a WAR/600 PA basis:

And here’s one for Starting Pitchers, on a WAR/200 IP basis:

Here’s one I haven’t shown in an article before: the average reliever aging curve on a WAR/60 IP basis:

For one year deals, we have no need for aging curves, just the player’s projection from Fangraphs Depth Charts. For multi-year deals, we can use these aging curves to adjust a player’s projection over several years.

So, using the player’s age, projection, and the above aging curves, I generated a projected WAR for each deal signed this year. I also added $10M to the player’s cost if he previously had a Qualifying offer extended to him, meaning the team signing him lost a draft pick. This was not added into Walker and Hellickson’s deals, as the player would need to reject the Qualifying Offer for compensation to be attached. It was added in though, for Justin Turner, Kenley Jansen, and Jose Bautista, because the Dodgers or Blue Jays would have eventually received a pick for each of those players had they not signed them themselves.

Using a 5% inflation rate, I then translated future wins into the equivalent value of today’s wins. I then totaled up the adjusted win totals, and divided it by the total money (and picks) spent on free agents this winter, to get a price per win on the free agent market this winter. $1.5B was spent to acquire 169 wins, coming to an average of $9M per win. Here’s a scatter plot:

The farther to the left of the line each dot is, the more the deal was an overpay according to the projections. The farther to the right of the line, the more the deal was a discount for the team. Of course, that’s just according to the projections. Teams can and do have their own evaluation tools, and when a team pays a player way more than the projections imply, that should be because that team sees that player as more valuable than the projections. What’s most important is the overall average, and the fact that the guaranteed dollars correlates well with projected value. The big cluster in the bottom left-hand corner are mostly all one and two year deals.

According to the projections, here are the ten best deals, using the $9M/WAR price:

I was driving the Justin Turner bus all the way until he signed, and was disappointed that he signed for so little for a team that wasn’t the Cardinals. Interestingly, the Kenley Jansen and Aroldis Chapman deals both came in cheaper that the projections expected, despite a general sentiment that the market over-values relievers. The projections think the Brewers’ bet on Eric Thames’ overseas numbers is a good one. Likewise for the Rangers’ bet on Carlos Gomez’s turn around near the end of the year. The Dodgers place three players on this list, which doesn’t seem fair when considering their ridiculous financial muscle.

If we extended this a little bit more, we’d find the Brett Cecil contract at a nice surplus of a few million:

His 2017 salary includes a $1M signing bonus. Cecil projects to be a minor surplus in the first two years, with the surplus in year three cancelled out by a deficit in year four. Not bad for a four year deal.

That’s the good news for Cardinals fans. Here’s the bad news, at least according to the projections. The ten worst deals:

Due to a projection of just 2.1 WAR (he’s averaged 4.0 the last two years) as well as a draft pick compensation penalty, Fowler’s deal ranks out badly. Only Ian Desmond’s signing grades out worse, due to a 1.5 WAR starting projection and a similar guarantee. I disagree with that projection strongly though, seeing more of a 3 WAR player in Fowler. The major points of difference are that I see less regression in Fowler’s BABIP (which is backed up by the Statcast data as well as being one of the faster guys in the league) and I believe more in his defensive transformation. Craig also found that players like Fowler age better than average. Zips originally had Fowler at 2.8 WAR in 2017, but I guess the inputs changed a little bit since then.

With that in mind, let’s look at how Fowler’s deal grades out according to the projections:

Jeez, I’m glad I don’t agree with the projections. They see only one year of Surplus Value. If you see him as more of a 3 WAR player, it ends up pretty close to break-even.

So what does all this mean? Generally, statistically-inclined baseball fans have quoted $8.5M as the price of WAR this winter. To be honest, a lot of players have went for less than I expected them to, leading me to believe the cost would come out less than that. However, we ended up a little higher. Several relievers going for less than projected was probably the most surprising aspect. Four made the ten worst deals though.

There’s a lot more ways to analyze this data, but we’re unfortunately out of time for today. In the future, we’ll look at this data in some more interesting ways. Still, we found some fun stuff today, and well, one not so fun thing. Hey, I for one am still glad they signed Fowler. More on Thursday.