This was a great read -- it's starting to make sense to me. The author admits that $/WAR isn't linear, which I agree with.
Where I disagree, the author mentioned that it doesn't really matter what number you use ($5M per, $7M per, $9M per, etc.), since it's all for establishing player values in relation to one another. As long as the number you pick is the number you use consistently, the author argues things are fine, because it's just expressing production in terms of $.
This is where I disagree.
Since the basis is to assess value, the metric is comparing whatever $/WAR figure they come up with against actual player salaries. If it were all hypothetical, I'd be fine with it ... But since half the equation is *real* salaries, it's important to that $/WAR accurately maps to real player salaries as well. Otherwise, half your equation is in Greek, and the other half in Mandarin.
This is really better expressed as a logarithmic curve than as a linear equation, but we just keep defaulting to linearly saying a player's value is $9M/WAR.
Let's read it like this:
Let's say JT Realmuto is a 4.5 WAR player for each of the next two seasons. So at $9M/WAR, that's worth $81M. Now, he has two arbitration years left, so let's say that equates to $20M of anticipated salary (random guess on my part, this is just an example). So that means he (or more importantly, his contract) has $61M of *value* across MLB.
But, as we've seen and discussed, and the author admitted ... $9M/WAR is not a clear, accepted, linear thing. This metric is really curved somehow. Trout won't get $70M+ annually. As 5 WAR players, Machado and Harper sure aren't getting $45M annually. And as the perfect example of a 1 WAR bench player, Charlie Culbertson sure wouldn't get $9M annually on the open market, unless someone REALLY believed in him enough to sign him as a starter.
So, what if we assume WAR is curved, something like this:
WAR1= $4M
WAR2= $5M
WAR3= $6M
WAR4= $8M
WAR5= $9M
WAR6= $4M
WAR+= $2M
Granted, I'm completely making this up off the top of my head ... But I still feel there's more reality in this than in $9M/WAR.
Under such a curved scale, your 1 WAR Culbertson would sign for $4M AAV on the FA market. Your 5 WAR Harper/Machado would sign for $32M AAV. Your 8 WAR Trout would sign for $40M AAV. And your 4.5 WAR Realmuto would sign for something like ~$27M.
While these numbers are obviously napkin-paper math, and very unscientific, I think that sounds reasonable. Or, at least like a reasonable approach at approximating value. I could see all of those contracts being signed at roughly those amounts. And given the STEEP dropoff in $/WAR when you get to Trout's level, I could see how teams would still trade a bounty for him in the future, even if he signs a $40M AAV deal ... As there would be *perceived* value, as the scale drops off so dramatically at the upper extreme.
Thus, if Realmuto's FA AAV worth is $27M annually, X2 years of control, that player's contract covers $54M in anticipated production. Since his arbitration will probably only cost ~$20M for those 2 years, he has ~$34M of value. Compare that to the $61M of value calculated above at $9M/WAR, and you can see why this matters ... Those numbers are highly different.
I believe that's why Anthopolous was wise not to overpay in prospects, and why the Phillies may regret doing so.