How To Shop In the Non-Tender Market… Successfully

I imagine that, for a front office exec, there’s nothing quite like the buzz you get from picking up another team’s non-tender and getting value from that player. Maybe it’s just ‘one man’s ceiling is another man’s floor,’ but in a business where one sector of the market has to continually work to find value in surprising places, it’s an important moment.

But is there much success to be found in the bargain bin? These are players that their own team has given up on — and we have some evidence that teams know more about their own players than the rest of the league, and that players that are re-signed are more successful. What can we learn from the successes and failures that we’ve seen in the past?

To answer that question, I loaded all the non-tendered players since 2007 into a database and looked at their pre- and post-non-tender numbers.

The first thing that leaps out at the page is that this is a tough thing to do. Since the 2007 season ended, 233 players have been non-tendered. That’s not counting this year’s 43. Of those 233, 20 were worth a win or more the season after they were let go. You’d have to invite twelve non-tenders to camp to get one of those guys — and be good enough to know in Spring Training which one was your guy.

There’s another way to tell that teams are generally pretty good at knowing which guys to let go. If you look at the non-tender list as a whole, 41% were above replacement the year they were non-tendered, and 27% were the next season. Not only did they pick players that didn’t get better, it looks like they were mostly right about which way those players were headed. Players normally get worse, this isn’t a group of 30-year-olds for the most part. In order to be non-tendered, you have to be under team control in the first place.

Eight non-tenders went on to have above-average seasons the next year: Willie Harris in 2008, Kelly Johnson, John Buck and Jack Cust in 2010, Russell Martin in 2011, Jeff Keppinger and Joe Saunders in 2012, and Mike Pelfrey last season. Just from looking at the success stories — particularly the number one former non-tender in the sample, Kelly Johnson (5.4 WAR) — you get a sense that you want your guy to have demonstrated defensive skills at interesting positions. Or, in the case of Keppinger and Harris (3.2 WAR, and second), the ability to play all over the diamond. Or, in the case of Cust, a singular skill at the plate.

So, the easiest way to spot a future contributor in the non-tenders? It’s so simple it’s stupid. Look at their value last year. Of the 20 players that contributed a win after being set free, only four were below replacement the year their team let them loose. That’s Matt Capps (1.2 WAR in 2010), Alfredo Aceves (1.1 WAR in 2011), Jonny Gomes (1.1 WAR in 2009) and Aaron Miles (1.8 WAR in 2008) in case you were wondering, or a platoon slugger, two relievers and an infielder that had one good year with the glove. Eight successful non-tender pickups put up between zero and one win before getting released, and the other eight were worth more than a win. So if you only look at non-tendered players that were above replacement, you double your odds of finding a useful player.

It’s still no better than a one-in-five shot if that’s your player pool. But it does say that Jerome Williams is probably a better acquisition than Dylan Axelrod, just by the numbers. Perhaps the usefulness of past non-tendered catchers makes J.P. Arencibia interesting, but Chris Nelson, coming off a below replacement level season and having struggled to show defensive value at any position — he’s got a major hill to climb until he’s productive for a major league team.

If you set your sights low — say, a fourth outfielder, a backup catcher, a super-utility guy, or a swing man between the bullpen and the rotation — you might find your guy in the non-tender list. You’ll want to look into players like Justin Turner, Ryan Webb, Jerome Williams and Derrick Robinson, probably. At least that way you’ve bettered your odds.

Here’s the dataset if you’d like to poke around.