Nice thread, he made a very convincing case backed up with solid data but the Rays' countered with bad-sounding terms like "Blown Saves" and "Meltdowns" that seem to have swayed the arbitrators.

This is a fascinating thread. It also seems like the panel placed a large value on his injury causing him to miss the playoffs, but they're deciding on his wage for this year, not his wage for last year.

Thompson had a nice list of comparables that were better compensated than he was.

The most important statistics for a middle reliever/set up man are holds and leverage index both of which I excelled in both the platform year and in my career with consistency....

The Rays did an excellent job discrediting holds and leverage while targeting me on blown saves, lack of LHH usage, and a fangraphs metric called “meltdowns”.

Well, yeah. Holds are a goofy stat and easy to discredit. Leverage index is a measure of how you're used. Did he really think these were the best stats to measure middle relievers or did he think they were the ones that made him look the best?

The inside look at the process was really interesting though.

It'd be nice if they had arbitrators who knew baseball, but at the very least, they should provide the players and the teams a summary of the arbitrators' knowledge. That way, they know going in if they are going to have to make a case for FIP over ERA or if they're going to have to explain where the pitcher stands.

Well, yeah. Holds are a goofy stat and easy to discredit.

Yes but if you're in a hold situation and fail that's a "blown save" even though you had no real opportunity for the save. If the hold is a BS stat (and it kinda is) then so is the "blown save" applied to a hold situation. For example, two of Thompson's "blown saves" came in the 6th inning. They might have been bad performances but that's a "blown hold" not a "blown save." (Or we start divvying up the one save among the 3-5 relievers that protected the lead ... which sounds like an interesting idea to me.)

Now it's true that none of this looks that flattering to Thompson. He pitched in 18 "save situations" with 10 holds, 3 saves and 5 blown "saves" so he did the job just 13 of 18 times which isn't anything special. In his case, 4 of those 5 "blown saves" were games he entered with nobody on base so pretty legit bad performances.

It's an interesting case in that he was Jekyll and Hyde. Through June 18, the date of his 5th blown save, he had an ERA over 6 (1-2, 3 sv, 3 holds, 5 fails). Over his last 22 innings/appearances of the season, he had an ERA of 1.25 (2-1, 7 holds, 0 fails). He can point to the amazing part, they can point to the disaster part.

Note, with the 3-batter rule, the hold is a bit more legit. No more come in, pitch to 1 batter, get credit for a hold (unless you come in with 2 outs). You can still get a hold for some pretty minimal, mediocre efforts but it's closer to capturing what it was supposed to.

EDIT: I hate Twitter formatting ... any chance somebody on the web has collated the relevant info into a more readable format?

Corbin Burnes was pretty unhappy with the Brewers arb case which per CB was that the Crew missed the playoffs because Burnes tailed off in the second half.

Supposedly Matt and Corbin met the next day to talk. But the Brewers didn’t dispute Burnes version. All to save $650k

The belief is that Mark tasked Matt to “win” his first major negotiation and Arnold took it so literally alienated Crew’s best player

Good work. Just awesome

On this Rays guy didn’t he have an agent doing his side or the case? Shouldn’t the agent know what to do and not do whatever the defined guidelines?

The inside look at the process was really interesting though.

My wife's uncle worked arbitration cases for several teams in the 80's. He still has some of his reports. I read one once, middle reliever for the Reds or Astros I think. He had excellent stats compared to many closers at the time, but it was all about the saves, how hard it is to get the 27th out, not too many people can do it... Typical BS, but he won the case on behalf of the team.

On this Rays guy didn’t he have an agent doing his side or the case? Shouldn’t the agent know what to do and not do whatever the defined guidelines?

I think his whole point is that this process seems arbitrary to the point where you don't know what to do and what not to do.

OTOH, doesn't seem all that different than trying a case before a jury of layman, you should probably expect to have to educate and rely on "buzzwords", and how does he represent a reliever and not know the Meltdown stat?

My wife's uncle worked arbitration cases for several teams in the 80's. He still has some of his reports. I read one once, middle reliever for the Reds or Astros I think. He had excellent stats compared to many closers at the time, but it was all about the saves, how hard it is to get the 27th out, not too many people can do it... Typical BS, but he won the case on behalf of the team.

Was the casual attitude of the arbitrators described in this article similar to what your wife's uncle observed?

On this Rays guy didn’t he have an agent doing his side or the case? Shouldn’t the agent know what to do and not do whatever the defined guidelines?

I think he did, or at least some type of advisor because he used "we" when describing his side.

Was the casual attitude of the arbitrators described in this article similar to what your wife's uncle observed?

Hard to say. I would suspect no. We haven't talked about it for 10 years or so.

Supposedly Matt and Corbin met the next day to talk. But the Brewers didn’t dispute Burnes version. All to save $650k

The belief is that Mark tasked Matt to “win” his first major negotiation and Arnold took it so literally alienated Crew’s best player

The scuttlebutt is that the Brewers front office believes they'll never be able to match whatever Burnes is going to get on the open market and he's gonna be gone as soon as he hits free agency, so it doesn't matter how he feels about the arb hearings. If the team doesn't win this year, Burnes is a tremendous chip to trade next winter.

The arbitrariness of arbitration (bad album title there) I suspect comes down to which stats are presented. For Thompson, the Rays use his blown saves, for another reliever they'll use his IRS, for another one they'll bang him on his FIP ... or his ERA if it's worse than his FIP. And the players are no doubt doing the same cherry-picking. It would be "fairer" if the arbitrators said "we will evaluate reliever cases on these 10 criteria so both sides need to address each criterion" or some such. If "blown saves" are important for Thompson then they're important for everybody and the middle relievers with few blown saves get paid. If holds are important then the guys with the most holds get the money.

They could remove a lot of the ill feelings generated by arbitration by issuing decisions based on briefs and not having a hearing. Give each side a chance to file a response to the other so you can argue why meltdowns are a junk stat.

There is nothing to gain by having a live hearing except racking up attorney fees. But there is something to lose if you see a representative of your team actively arguing that you're not such a great player.

Well duh, the main link goes to a news article that does what I wanted. Hooray for me but somebody coulda told me. :-)

I agree on the transparency argument at least a bit. At one job I had, in a fairly large organization, I got a very strong performance review from my boss so was looking forward to a raise.

But then the bosses get together and review all the folks at a given level. This is supposed to ensure that each of the bosses is applying similar criteria in their evaluations -- that I did as much to get my 5 out of 5 as the other person who got 5 out of 5 and that I did more than the one who got 4 out of 5, etc. All seems perfectly sensible ... except I'm not in that process and had no means of knowing how that decision was made.

So I questioned it when my raise wasn't sufficient. I only went to my boss and said "is this process documented anywhere? I know the criteria by which you are evaluating me but by what criteria am I being evaluated by people I'm not working for? All I can do is satisfy your criteria (5 out of 5), how can I also satisfy the unknown criteria of someone I don't work for?" I didn't say anything about "give me a bigger raise," I just asked "what are the criteria to get a bigger raise?" The next day I was given my bigger raise.

I suspect that if the WAR formula for pre-arb players continues without problems, they'll eventually use that to replace arbitration.

Thompson's bWAR is the worst of his list of comparables. Must be why his team focused on LI.

We had to assume that the arbitrators were savvy enough to understand basic rules and statistics. I believe that assumption was incorrect.
...
The most important statistics for a middle reliever/set up man are holds and leverage index both of which I excelled in both the platform year and in my career with consistency.

Flawed arbitration strategy. Arbitrators are not professional sabermetricians, they are professional arbitrators. The case you want to make is "This is what similar players make, taking service time into account and adjusting for growth in salaries."

Saying that the most important stats for a middle reliever are holds and/or leverage index is an opinion. Some sabermetricians might hold it, others might disagree, none of them are in charge of the MLB salary structure. You're asking the arbitrator to say that your analysis is better than the MLB salary structure, which they are going to be very reluctant to do.

Thompson's bWAR is the worst of his list of comparables. Must be why his team focused on LI.

Not looking it up, but in this instance (relief pitchers) I rate fWar as the better animal, although in reality a modified war using runs created for the runs allowed formula might even be more accurate. I like bWar for starting pitchers, for relievers a war formula using runs created, inherited runners/scored and leverage would probably be a better formula for their evaluation (and I get you can't really have different formulas for players based upon their role/position, but for an evaluation of how good they are relative to their peers in the role, it makes sense)

Bill James had an excellent article about the arbitration process in the '80s. One of the questions he was asked was whether he used his own stats in consulting for arbitration cases. He said he didn't do that often, partly because you've only got so much time to make your case, but mostly because arbitration is really about fitting the player into the existing salary structure (the salary structure, not the value structure).

If teams are paying Proven Closers for saves, you're not going to have much success arguing that holds are a better metric for middle relievers.

#19: While I think I get (and agree with) the overall point, I'm not quite seeing what you're saying. To claim you are "similar to" Graterol and better than Staumont, Bedrosian and Hernandez (and not comparing yourself to whoever the stud middle relievers are these days), you have to do that on some criteria. The hold is an "official" MLB stat and is the most obvious counting stat for middle relievers. Leverage index speaks to usage (garbage time vs important time), an important area of comparison for relievers I would think. The Rays meanwhile pulled out "meltdowns."

"My leverage index shows I am the same type of middle reliever as Graterol (Thompson's was in fact slightly higher) and my equal number of holds show I was as successful. Graterol makes $1.225, I therefore deserve my request of $1.2"

LI is a great way to show how much the team trusted him with leads and pressure situations. Holds and blown saves show how often he did the job. Last year 3 saves, 10 holds, 5 blown saves. 18 times given a lead, lost it 5 times. Two of his 5 blown the team came back to win in the end. For an emotional argument I'd be going in that direction - 18 times handed a lead, lost it only 5 times two of those times the team was able to win anyways, thus a 15-3 record when he was given a lead - the most important stat one could argue (if facing people who know little then even better to use). 2 of his blown saves that ended in losses he only gave up unearned runs (ie: not his fault).

Basically I see that his agent went for a 'we are facing arbitrators who know baseball' when he should've gone with 'these guys know nothing, lets hit them with stuff that sounds good', or have both plans ready and judge which is best when you start talking by how they react to your opening - mix some stats and some emotion and see how they move around, then go with plan A or B depending. Ideally do a mix of both, so you have your bases covered. Have charts (if allowed) to show player A/B/C which highlights how much better he is than A and B who make more than the midpoint. There is a reason politicians use big signs to get points across - they know their audience probably knows little so making stuff look good is #1, even if it is all bs.

Funny, I'm a stats addict and had no idea about 'meltdowns' - From FanGraphs: WPA <= -0.06, Shutdown if WPA >= 0.06. Interesting way to measure how a guy did.

So Thompson had 17 'shutdowns' and 15 'meltdowns' which doesn't sound too good imo.
Graterol had 17 'shutdowns' and 11 'meltdowns' last year.

So using that measure Thompson was a bit worse - 4 more meltdowns, but same shutdowns. A bit arbitrary as to what it is, and in the future his agent should damn well know what that is and be ready for it.

If you go before an arbitrator and he/she doesn't laugh in the face of your opponent who trumpets Blown Saves as a core argument - well, you're doomed.

as noted above, a reliever pitching in the 21st century before the 9th inning can't 'win.' if he protects the lead, he still can't get a save. if he doesn't, he gets a blown save.

"Holds" were, I think, an honest attempt at a workaround. but it doesn't work that much better than "Saves" do. too often the manager winds up pitching to the stat instead of trying to make his best efforts to win the damn game.

there was a Yankees postseason game BITD where manager Joe Torre - that Neanderthal ! - brought in Rivera in a brutal spot. don't recall if it was 2nd and 3rd, nobody out, 1-run lead, or the exact details.

but Rivera did an excellent job and iirc the Yankees won the game - but sorry, he got charged with a "Blown Save."

the Marlins had a manager years ago who was completely addicted to the Save Rule. so a couple of times, he brought in his "closer" with a 5-run lead, 2-outs, bases loaded in the 9th because that would score his guy a save.

meaning, he could allow a 500-foot grand slam, a double, a couple of walks, and then a rocket fly ball caught at the warning track - no worries, ballgame over, it's a save!

and if that meant he wouldn't pitch in a 9th-inning jam in the next game because he wasn't going to be asked to pitch for a third straight day - who cares?

all bow to the Save Rule

#22 -- I agree with you completely. I'm not endorsing blown saves as a good arbitration strategy, either!

On the other hand, given his list of comparables, I think the team wins this handily.

Ryan Thompson debuted in 2020.

Josh Staumont debuted in 2019.
Brusdar Graterol debuted in 2019.
Cam Bedrosian debuted in 2014.
Jonathan Hernandez debuted in 2019.

Service time is the biggest driver of salaries in non free agents, and all of Thompson's comparables have more service time than he does. One of them has way more.

Graterol debuted in 2019 but has 2.167 service time, a super-2. Thompson has a full 3 years service time. Thompson wins that one from the arbitrator's perspective. Staumont has 72 more service days, Hernandez has 41. So these guys are all equal. Bedrosian is a weird comp but back in his first arb year (way back in 2018) he made $1.1 after 138 IP of 101 ERA+.

For those that don't know, at the top of a current player's page, b-r gives his service time (sometimes you have to click "more info" first). This isn't updated in real time but every Nov 1 (or whenever). And from an arb perspective, it's service time not "years" that matter.

If anything, it sounds like Thompson and his rep were better prepared for the way arb is supposed to work -- here's how our guy performed, he's similar to these guys and better than these guys and the good guys get $1.2 and the worse guys get $1 so our guy gets $1.2. The Rays went with "but sometimes he sucks and when he sucks, he really sucks. And he got hurt at the wrong time."

On the other hand, he opened himself up to that argument because it was true in the first half including a pretty amazing stretch from May 6 to June 18 when he threw 11.2 IP, 21 H, 17 R, 15 ER -- he's lucky he still had a job. The rest of the year he gave up just 3 ER.

Graterol debuted in 2019 but has 2.167 service time, a super-2. Thompson has a full 3 years service time. Thompson wins that one from the arbitrator's perspective.

That’s pretty much Thompson’s argument, that Graterol is his closest comparator, and he has at least demonstrated that he is above the midpoint between his submission and that of the Rays, which should be enough to prevail. The other comparators are more of a ‘better than them, too’, argument. Thompson may be right, but the difference is small - they’re haggling about $200K. The Rays also took Colin Poche to arbitration over a $125K difference. Presumably the Rays think they come out ahead by nickel & diming their players during their years under team control, but I’m somewhat skeptical. Losing 1 player they’d like to keep might well outweigh what they won by contesting such small amounts.

I agree that service time is a better comparison than debut date -- thanks for looking that up. In that case, the list of comps doesn't seem that bad.

Thompson's tweet thread compares his career stats to several other players. He doesn't mention his platform year or how that compared to the platform years of others. And since Thompson's platform year had a negative bWAR, there could well be other reasons than arbitrator incompetence for the final decision.

Well, WAR (at least bWAR) needs to adjust for the unearned runner in extra innings. He's getting dinged on a 5.06 RA9 but, take out the Manfred Men, and it's 4.43. Most of those guys score and remember that the league average RA9 in extras is now over 9. His extra innings performance this year was a bit better than average (he gave up just 3 R in 3 IP!!) but it's being compared mainly to innings where the average RA9 is 4.whatever ... while they are charged as unearned runs, from the reliever's perspective, the unearned runner is clearly an inherited runner.

Now whether WAR for relievers needs to do something about IRs (which would require tracking base-out situations) is a good question.

And guess what. On Apr 12, hosting Oak, he came on in the 10th and gave up the unearned runner. WPA (at b-r) of -0.075 ... a "meltdown" although it was technically speaking a slightly above-average performance. Another of his "meltdowns" was also an extras unearned runner (but at least he did nothing right in that one either) as was one of his "shutdowns" (but that seems OK).

All of these reliever stats need to adjust for the extras silliness. While at a team/league level, it all comes out in the wash, the individual relievers who actually pitch in extras are getting penalized (and the ones who don't are getting a tiny reward). If teams are going to be pulling WAR and meltdowns into arb hearings, b-r and fangraphs have a responsibility to make them as fair as possible.

Whoa!! Apologies to b-r. There is a RA9extras factor in WAR now. I don't know if they've got it right but they won't be far off so Thompson's WAR should be about right.

EDIT: Some league numbers to work with

1 - 9: 4.28 RA9
extra: 10.13 RA9