Yup, it's fWAR. I get that it's a fantasy, in that Betts would not get a 1/$61m offer were he on the market.
But the pool of just Free Agent players tend to get about $8m per projected WAR, and that's where those numbers come from. Remember that these are averages, and the average excludes pre-FA players and their salaries. So sometimes David Ortiz gets paid $16m to produce what Fangraphs thinks is $36m worth of WAR — amazing value! Sometimes Mitch Moreland gets paid $6.5m and produces $5.5m worth of WAR, and we got roughly what we paid for. Sometimes Pablo Sandoval gets $95m and produces -$8m worth of production — a total nightmare. It all goes into the average. These numbers used to rise steadily a few percent a year, but these days they seem to have plateaued.
Because they are averages, these $/WAR numbers do get a bit wonky at the extremes. Trout, for example, is an amazing value at a $36 AAV. If you, like me, think that Trout is likely to exceed 4.5 fWAR next season, he should produce a fair amount of surplus value for the Angels even at that steep salary. He is the best player, the best-paid player, and yet a surprisingly decent $/WAR value among players with 6+ years of service time. In my view, this is itself an argument for extending Betts even at huge dollars. Even if we pay him something outlandish like 8/$300 or 10/$350m, if he averages 5 fWAR over the life of the contract (a high bar, but not crazy for a player who has averaged 7.5 fWAR/162 games for his career), it would be a good FA value. That's why he's valuable. His market value — the money he can command through arb and eventually FA — is unlikely to match the value of his projected production (health allowing), simply because contracts don't go high enough. So yes, I know that Betts won't get paid $60m a year, and in that sense he isn't "worth" that amount. But that recognition itself helps make clear why he's a desirable player: he'll get paid ~$28m to produce an amount of WAR that would normally cost $61m in free agency.
And you're of course right that these values don't help explain what actual rosters look like, but only because real rosters aren't entirely made of free agents! An all FA team, with no pre-arb or arb players to get extra value, and average FA return would get ~25 WAR for $200m, and go 67-95. What that tells us is that you need to get surplus value out of your artificially cheap pre-arb and arb players to put a good team together. Even more than you want Mookie Betts to give you a 6.6 fWAR season for $20m, as he just did, you want a Rafael Devers to give you a 5.9 fWAR season for $500k.
What these $/fWAR surplus value calculations are really good for is comparing trades. The Fangraphs guys have now put $ values on tiers of prospects by projected future value, which makes it possible to estimate how good a prospect you can expect to get back for a given veteran.
FWIW, I started this thread and I agree with every word of this post. I don't think any of these trades are as smart as giving something like the 2018-19 roster another spin in the hope that Sale, Price and Eovaldi can each make 25+ healthy starts. Hell, we've even developed a bullpen since April 2019.
If you read my initial post — which I don't mean to suggest you didn't — I don't actually endorse any of these trades. I actually think all of them would be bad ideas.
I can squint and imagine it making sense to deal JD Martinez to Chicago, if they will actually send us Reynaldo Lopez in return to be our fifth starter. Then we could either find a cheaper positionless slugger — Eric Thames? — or else promote Bobby Dalbec. If we then spent some or all of the saved money upgrading the right side of the infield, we might even be a better team after such a trade. That I think it would actually make sense to deal Martinez to Chicago for Lopez suggests to me that Chicago would not accept such a deal.
(The Trade Simulator also suggests that Martinez-Lopez is extremely tilted towards the Red Sox, but that maybe we could get CWS' #4 prospect Dane Dunning, a 23 y/o RHSP coming off TJS after a very nice 2018 between high-A and AA, and buy low on their #8 prospect, The Good Basabe, after a so-so season in AA. If we actually needed to shed payroll, I might do that deal. It hurts us in 2020, but adds a decent CF prospect and a good SP prospect to the high minors.)