Few Marginal Players Will Get Qualifying Offers Today
Today at 5 p.m. E.T. is the deadline for teams to make a qualifying offer (QO) to their free agents. For years, this decision was straightforward: teams handed offers to anyone remotely worth it. Why not? Every player rejected them. It was a free draft pick.
Meanwhile, evidence mounted that players should adjust their strategy of “always reject.” Nelson Cruz, Stephen Drew, and Kendrys Morales (among others) showed that rejecting the offer wasn’t always a good idea. These players cost themselves big time.
Can you imagine the frustration they felt at reaching free agency only to see a big payday slip through their grasp? This year’s free agents are watching and learning. And teams are watching and learning with them.
When these guys rejected the QO, their price tags increased by about $8.1 million. Research by Matthew Murphy at The Hardball Times shows that, prior to 2014, a draft pick in the compensation range (typically picks No. 26-30) was worth this much, or a little more than a win. Cruz, Drew, and Morales gambled that teams would pay this additional price for them.
They lost their bets. Teams priced the draft pick into their assessments of each player and found the math upside-down. Because the teams had all the leverage, the players had to concede some salary.
Last year, three players got wise and adjusted their approach. In the 2015-2016 offseason, Brett Anderson, Colby Rasmus, and Matt Wieters accepted QOs from the Dodgers, Astros, and Orioles, respectively. Meanwhile, Ian Desmond became the latest example of a player weighed down by the draft pick he incurred by rejecting the QO.
The signal became clear: marginal players shouldn’t reject the QO. If they do, they’ll get burned. This year, the signal will manifest itself as a higher probability that a marginal player will accept a QO. I expect teams will read this intent and, accordingly, be more selective this time when making their QOs. Fewer marginal players will likely get one.
To model the break-even point teams need to justify a QO, I estimated what each player might be worth on a one-year deal and compared that to the $17.2 million guarantee they’d get if tendered a QO. I used the following inputs to compute break-even points for many free agents:
- A reasonable one-year contract offer. To get this I divided the total value of multi-year contracts, as estimated by the FanGraphs crowdsourcing project, in half.
- The value of a QO: $17.2 million.
- The value of a compensating draft pick: $8.91 million.
The formula is:
Break-Even % = 100 – (100 * ((QO Value – One Year Contract Value) / Draft Pick Value)
where Break-Even % is between 0 and 100.
Consider Luis Valbuena. No one thinks Valbuena is worth $17.2 million per year, but he’s a useful player who should land a multi-year contract as a free agent. In years past, an aggressive team would’ve made him a QO in order to get a draft pick. In the soon-to-be published results of this year’s contract crowdsourcing exercise, readers estimated Valbuena’s contract this winter (without a QO) at $25.6 million, so I estimated his one-year value at half that: $12.8 million. In reality, he’ll probably sign for less than that on a per-year basis, but he’ll probably get two or three years, so that’s my guess at what his open-market value would look like if he only wanted a one-year deal.
Using the formula noted above, Valbuena’s break-even point is calculated as follows:
100 – (100 * ((17.2 – 12.8) / 8.91)) = 100 – 49 = 51%.
If the Astros though there was less than a 51% chance that Valbuena would take the QO, they could justify making him the offer. The potential value of a draft pick makes up for the risk of overpaying him by a few million if he takes the one year deal. However, I’m confident Valbuena is more than 51% likely to accept a QO this year. Last year or the year before, maybe not. But he knows that if he rejects the offer, the draft pick will hurt him badly. He’ll have to convince other teams he’s worth something like $20 million for one year or $35 million over two, since they’ll be factoring in the tax they’ll have to pay to sign him as a free agent.
Therefore, the Astros shouldn’t make him an offer, even with the lure of a draft pick if he turned it down and signed elsewhere. If they do, they’ll likely find themselves paying $17.2 million, after he and his agent correctly ascertain that the QO would crush his market value, and a multi-year deal probably isn’t available on the open market with a draft pick attached. Also, the Astros would have the additional headache of finding playing time for Valbuena while also having to work Alex Bregman and Yulieski Gurriel in at the same positions Valbuena plays.
Rasmus is another player who won’t get a QO, even though he got one last year. His break-even point is 47%. If he’s any more likely to accept a QO, the Astros shouldn’t make it. With an estimated one-year contract of $12.5 million, Rasmus is more like 99% probable to take the QO.
The last interesting guy to examine is Brandon Moss. His break-even point is 41%. Like Valbuena and Rasmus, Moss would take the QO in a heartbeat. This may be why the Cardinals reportedly won’t make him one.
The following table shows other eligible free agents, sorted by Marginality, which is 50 – (|50 – Break-Even %|). The closer to 50% a player’s break-even point is, the more marginal the player is. (Once again, data taken from contract crowdsourcing effort.)
| Player | Marginality | Est. Total Contract* | Est. One-Year Contract* | Break-Even Point (%) |
|---|---|---|---|---|
| Luis Valbuena | 49 | 25.6 | 12.8 | 51 |
| Colby Rasmus | 47 | 25.1 | 12.5 | 47 |
| Brandon Moss | 41 | 23.8 | 11.9 | 41 |
| Edinson Volquez | 28 | 29.4 | 14.7 | 72 |
| Michael Saunders | 24 | 30.3 | 15.1 | 76 |
| Mike Napoli | 24 | 20.9 | 10.4 | 24 |
| Matt Holliday | 21 | 20.5 | 10.2 | 21 |
| Doug Fister | 15 | 19.2 | 9.6 | 15 |
| Brett Cecil | 3 | 17.3 | 8.6 | 3 |
| Angel Pagan | 1 | 16.7 | 8.4 | 1 |
| Edwin Encarnacion | 0 | 90.4 | 45.2 | 100 |
| Kenley Jansen | 0 | 77.1 | 38.6 | 100 |
| Justin Turner | 0 | 70.9 | 35.5 | 100 |
| Ian Desmond | 0 | 62.7 | 31.4 | 100 |
| Dexter Fowler | 0 | 61.6 | 30.8 | 100 |
| Mark Trumbo | 0 | 59.2 | 29.6 | 100 |
| Jose Bautista | 0 | 65.3 | 32.7 | 100 |
| Jeremy Hellickson | 0 | 54.1 | 27.0 | 100 |
| Neil Walker | 0 | 50.5 | 25.3 | 100 |
| Wilson Ramos | 0 | 43.9 | 21.9 | 100 |
| Matt Wieters | 0 | 40.3 | 20.1 | 100 |
| Jon Jay | 0 | 14.0 | 7.0 | 0 |
| Sergio Romo | 0 | 13.6 | 6.8 | 0 |
| Joe Blanton | 0 | 13.1 | 6.5 | 0 |
| Matt Joyce | 0 | 12.0 | 6.0 | 0 |
| Bartolo Colon | 0 | 14.7 | 7.3 | 0 |
| Chase Utley | 0 | 12.2 | 6.1 | 0 |
| R.A. Dickey | 0 | 11.6 | 5.8 | 0 |
| Koji Uehara | 0 | 8.5 | 4.2 | 0 |
| Rajai Davis | 0 | 11.5 | 5.8 | 0 |
| Pedro Alvarez | 0 | 10.9 | 5.4 | 0 |
| Colby Lewis | 0 | 9.8 | 4.9 | 0 |
| Santiago Casilla | 0 | 8.5 | 4.2 | 0 |
| Greg Holland | 0 | 8.2 | 4.1 | 0 |
| Jonathan Papelbon | 0 | 7.3 | 3.6 | 0 |
| C.J. Wilson | 0 | 7.1 | 3.6 | 0 |
| Kurt Suzuki | 0 | 7.7 | 3.8 | 0 |
| Austin Jackson | 0 | 7.5 | 3.7 | 0 |
Valbuena, Rasmus, and Moss are the players most on the bubble. The next tier includes Edinson Volquez, Michael Saunders, Mike Napoli, and Matt Holliday. Their respective break-even points are 72%, 76%, 24%, and 21%. If you’re Dayton Moore, do you think Volquez is more than 72% likely to accept a QO? If so, don’t make him one. As for Napoli and Holliday, I guarantee you they’re more than 24% and 21% likely to accept a QO. Those thresholds are very low.
The guys who definitely won’t take a one-year, $17.2 million offer appear in the middle of the table. Their break-even point is at 100%, reflecting the fact that making guys like Justin Turner a QO is never a mistake. These guys will likely get offers. Because the decision to make an offer is so easy, their Marginality is 0.
Last come the guys who would take one-year, $17.2 million every single time. Making these guys a QO is always a mistake. A 0% break-even point for C.J. Wilson means if he’s even 0.000001% likely to accept a QO, Billy Eppler shouldn’t make him one. This decision is just as easy as deciding whether to make an offer to Turner, so Wilson’s Marginality is also 0.
If the physical game of baseball is one of constant adjustments, so is the business game. Last year, three players saw evidence of how rejecting a QO could hurt their income. They adjusted their business strategy and took the QO. This year, teams can expect more marginal players to do the same. Therefore, the teams themselves will adjust. The result will be fewer (if any) QOs to marginal players.
Ryan enjoys characterizing that elusive line between luck and skill in baseball. For more, subscribe to his articles and follow him on Twitter.
Any reason Kendrys Morales doesn’t make the table? I was wondering how marginal he was.
Only because he wasn’t in the list of crowdsourced contracts. I suspect he’s not very marginal. I’d guess he could get maybe 1/$7M on the market. That puts his marginality at 0. For him to approach a marginality of 50 like Valbuena, you’d have to think he could get 1/$12M, which I don’t think anyone would offer him.
You think he will be offered less than 1/$12 on the open market?
I think that, yes. But 1/$12M just gets him to ‘marginal’ status. And as I argued in the article, that means no QO.
To get Morales a QO you need to make the case he can get at least 1/$17.5M on the open market — thus leading to him rejecting the offer, thus leading to KC offering it.
I don’t think he thinks he can get this much on the open market, meaning he would accept a QO. Meaning KC won’t offer it.
Heyman reports the Royals will not extend Morales a QO, so I guess he’s pretty marginal
Shocked that Volquez is that high. It doesn’t sound like they plan to extend one, which I think is for the best.
Maybe I am misunderstanding but I read the above table and explanation in the article to say that (for example) a guy with a 0% chance of accepting the one year, $17.2 million qualifying offer is Matt Weiters?
Do you really think so?
Walker is another one that I think isn’t a slam dunk to reject the QO although it should be noted that he’s coming off season-ending back surgery, which the model can’t account for, as noted with Ramos’s (much more serious) injury, and that affects his value negatively, which in turn makes him more likely to accept the QO, and again the model doesn’t account for this.
But Weiters…yes, he’s the #1 (uninjured) catcher on the free agent market. So like a guy such as Hellickson w/r/t pitchers, he’ll benefit from weak alternatives for teams in need of help.
But it might be noted that the Orioles couldn’t have agreed with your model no? Because if they did, they would have offered Weiters the QO, certain that he would reject and they’d collect the draft pick.
Unless I’m misunderstanding, which as noted above is a distinct possibility.
“Maybe I am misunderstanding but I read the above table and explanation in the article to say that (for example) a guy with a 0% chance of accepting the one year, $17.2 million qualifying offer is Matt Weiters?”
No. The 0 for Wieters doesn’t refer to him 0% likely to accept the offer. It refers to the fact that the decision on whether to make him an offer is a slam dunk. The 0 is Wieters’ ‘Marginality’. Higher marginality (Valbuena, etc) means the decision to offer the QO is harder because the break-even % is closer to a toss-up (50%). Lower marginality (Wieters et al) means the decision to offer is easier.
“But it might be noted that the Orioles couldn’t have agreed with your model no?”
Noted. Hey, I’m sure smarter folks than me work in Baltimore’s front office 🙂 They also have more knowledge about his habits and tendencies than I do. It seems they felt he’d accept an offer, so they didn’t make one. I don’t doubt they made a reasoned assessment.
Thanks for explaining Ryan.
Very interesting article, by the way.
Wilson Ramos broke the model here because of his injury. On the table, he is “100%”, which without the injury would be absolutely correct. But the Nationals announced the are NOT giving him a QO, because getting $17.2 mil for a one-year rehab is a no-brainer YES decision for Ramos, and not such a good deal for Nats
It’s interesting that you say that, because the ACL tear happened a few weeks prior to the crowdsourcing exercise which estimated a total $44m deal for him. That’s why I didn’t discuss him in the article; knowledge of his injury was baked into the estimate.
We’ll see if the Nationals end up regretting that decision. I’m sure Ramos is happy as he won’t have a draft pick tied to his value.
The crowd thinks Ramos is getting $44M? Consider me floored, and I LIKE Ramos.
What in the world are these 1-year estimates? Encarnacion at 45MM for a year? That’s beyond absurd. I don’t know what his crowdsource is, but if it’s like 6/90 and you just divided by 2….. C’mon Man.
That is indeed what they did. I think the logic isn’t that any team would actually pay Encarnacion that much on a one-year deal, but that that is the offer it would require for him to forgo a multi-year contract. Yes, the methodology is clumsy, but it makes a little bit of sense.
Except that number is completely meaningless for the breakeven % calculation- the number that matters from the team perspective is the VALUE he’s expected to provide in the year covered by the QO,
Why are you viewing this exclusively from the team perspective?
The player gets to choose whether or not to accept the QO. And the number you’re focusing on is related to the player’s perspective, not the team’s.
Since EE would need $45m to “break even”…it doesn’t make sense for him to accept a $17m QO.
Back to the team perspective…This info means it’s a no-brainer to extend a QO to EE. Because he either accepts and you get him$17m (a steal), or you get a draft pick.
Because the article contains a calculation of a breakeven % FROM THE TEAM’S PERSPECTIVE. The player’s risk tolerance/time preferences have nothing to do with that calculation.
They do though. If a team knows that a player is going to refuse the qualifying offer, they then know it is safe to offer one, whether they want that player back or not.
The perspectives are one and the same. When deciding whether to offer the QO, the team should base their perspective on the player’s perspective.
I originally did this exercise using WAR. It doesn’t change how the top and bottom players are perceived. Turner will always get an offer and Jon Jay will always not get one.
But using WAR identifies the wrong players as being marginal. For example Trumbo (1.1 projected WAR next year) is marginal by WAR method. In reality he’s not marginal at all. FanGraphs readers projected he’ll get a contract worth $60M in total. Even if you think Trumbo will get $50M or $40M, with a draft pick attached to his name, those figures are so much higher than $17.2M that he’s a lock to reject a QO.
Therefore, the Orioles should offer him one.
How the hell can a player have a breakeven accept percentage? Are you even trying to make sense?
Trumbo is a good example of why your calculation is completely wrong. Let’s stipulate that Baltimore agrees with projections on a 1.1-1.2 WAR contribution next season, well short of 17.2MM value. If they QO Trumbo, and he accepts, Baltimore just flushed money down the toilet by paying 17.2MM for a win. Therefore Baltimore’s breakeven % for Trumbo accepting is NOT 100%, and it’s nowhere close to 100% because they get creamed every time he accepts. They need him to decline around half the time to come out even on the offer.
And they know he will decline, because they know he knows he can get a multi-year deal that will pay him 3 times as much overall. He won’t take a one-year deal because he’s more likely than not to be less valuable next year at this time, and he knows it, and the Orioles know it.
Come on. This isn’t that hard.
That’s completely irrelevant to the breakeven% calculation. Come on. This isn’t that hard.
Ok, I can *maybe* see that. But that’s such a hand-wavy kind of calculation that I think it requires a lot more explanation in the article than “I divided X by 2 to come up with a reasonable one-year offer.”
I have no clue why you’re getting downvotes. Makes me think we’re just missing something. I guess Giancarlo Stanton is worth $160 on a one-year deal….
Same reason the people posting complete nonsense in response are getting upvotes lol
The formula is wrong, and leads to incorrect conclusions, like that a team should not offer C.J. Wilson a QO and give up a pick worth $8.91M even if he is almost 0% likely to accept the offer.
The formula should be:
Break-Even % = 100 – (100 * ((QO Value – One Year Contract Value) / (QO Value – One Year Contract Value + Draft Pick Value))))
The Break-Even for C.J. Wilson is about 40%.
The 40% of the time he accepts, the team will overpay him by about $13.6M ($17.2M – $3.6M)
The other 60% of the time the team will get a pick worth $8.91M.
(.4 * -13.6) + (.6 * 8.91) ~= 0
The Break-Even for Luis Valbuena is about 67%.
Even this is probably incorrect, shouldn’t we be using projected value and not projected contracts prorated to one year. I.e Valbuena projects for 1.5 war and 13.5 mil of value, more than the crowd projections.
rpeltz’s equation is right (and Ryan’s is wrong), though it might be simpler to express it as:
100 X Draft Pick Value / (Draft Pick Value + QO – One Year Value)
What you’re worrying about is a different issue: how to estimate the one year value. Ryan’s method of halving the multi-year contract estimate will probably work well enough for marginal players, those around the 1.5 WAR mark, because you expect something like 1.5, 1.0, 0.5 on a 3-year contract, with the first year providing half the total value. For better players it won’t work – e.g. I really doubt Encarnacion would get $45M on a 1-year contract, and your method would work better here – but they’re an easy decision anyway.
Can you say that a comp pick on average turns out to be a prospect in the 90-120th range? if the comp pick is around 30th and you consider that most good talents are around 3 years in MLB you could expect that comp pick statistically to become a bottom Tier top100 prospect.
that is quite a Price for a marginal guy to sign.
No, you can’t say that. The comp. round has started around pick #25 in the last two years. That’s because as teams sign free agents w/draft picks attached, they lose their first round picks, unless they are one of the first 10 teams to pick, in which case they lose picks from subsequent rounds. The picks don’t transfer; they just disappear. Typically this process shortens the 1st round from 30 picks to about 25, at which point the comp. round starts.
Here’s the 2015 draft order: http://m.mlb.com/draft/2015/order. Note the comp. round starts with the 27th pick (even though it’s not numbered that way).
As for the $8.91M, that’s not directly paid to a player, but it’s in the formula as a penalty to a team who decides to not offer a QO to a player. If a team doesn’t make an offer they lose the chance at gaining the pick, thus they lose a chance at $8.91M.
For teams like the Dodgers, if they don’t make an offer to Turner, they lose 100% of this value because Turner’s guaranteed to reject the QO, thus guaranteed to lose the Dodgers a draft pick. But they lose 0% of the pick value by not making an offer to Utley, since Utley will almost certainly accept the offer (so the Dodgers have no chance at getting this draft pick anyway).
Cespedes is also a free agent after opting out, and I assume should be in that break-even rate of 100% group.
Yes. He wasn’t in the list of crowdsourced contracts until yesterday, because he hadn’t officially opted out. But yes he is a lock to reject the QO, so the Mets made the correct move in offering him one.
I’m completely lost on the “1/2 of the predicted contract value” thing, apparently without any regard for the length of the predicted contract.
So to be clear, if the readers predict a 1-year, $12M contract for Mike Napoli, you assume he’s going to get a 1-year, $6M contract? And if they predict a 5-year, $125M contract for Cespedes, you assume he’s going to get a 1-year, $62.5M deal?
I’ll try to lay out what Hareeb and rpeltz are saying in a bit more detail. There are two questions here:
(1) What is the maximum probability of a player accepting a QO at which it makes sense for the team to make the QO? This is the “break-even” percentage.
(2) What is the actual probability of the player accepting the offer if made?
(1) The first question is relatively easy: if a player declines then the team gains the draft pick value, DPV ($8.9M), if he accepts then the team loses the QO ($17.2M) but gains the player’s on-field value (OFV, which is WAR x $/WAR ). So with a probability of acceptance of p the expected value to the team of making the offer is:
p X (OFV – QO) + (1-p) X DPV
The team wants this to be >= 0. If the on-field value is greater than the QO then this will always be true and the break-even probability will be 1, i.e. always make the offer. If not we want to find
p X (OFV – QO) + (1-p) X DPV >= 0
p X (OFV – QO -DPV) + DPV >= 0
p X (OFV – QO – DPV) >= -DPV
p X (DPV + QO – OFV) <= DPV
p <= DPV / (DPV + QO – OFV)
So for C.J. Wilson if we say, for the sake of the example, that he's projected for 1 WAR and the $/WAR is $8M, giving an on-field value of $8M, we get:
p <= 8.9 / (8.9 + 17.2 – 8M)
p <= 0.49
i.e. making a QO to Wilson makes sense if the team thinks the chance of him accepting is 49% or less. The problem with making a QO to Wilson is not, as Ryan argues, that it makes no sense even if he's practically certain to reject it: the problem is he's almost certain to accept.
(2) The second question doesn't have such a simple analytical solution – it's going to involve what the player expects to be offered, what he thinks he might get offered a year later (which also depends on whether he thinks he'll get a QO next year too), his tolerance for risk, and intangibles like comfort with the team, desire to play for a contender etc. Here I think Ryan is mostly right: players like Valbuena are very likely to accept because the difference between the QO and the they might get with draft-pick compensation attached is small. So Valbuena's probability of accepting is likely higher than his roughly 65% break-even point. Better players are less likely to accept, but there's not much room between Valbuena's 1.5 WAR projection and the roughly 2 WAR point at which the team doesn't mind the player accepting anyway – there aren't going to be many players in that window.