Twins Mailbag: Kohl Stewart, Mauer Extension, DH Options

Rooker hasnt seen AA ball yet, has he? And only has 40 games outside of rookie ball?

I'd go year to year on Mauer, unless they go after Santana which I don't think they do.

"Upon request by a Major League player to the Commissioner, the Commissioner may, on such conditions as he stipulates, consent to such player accepting a renewal of his contract at a salary rate less than eighty (80) percent of the rate stipulated for the preceding year, providing the salary rate is mutually agreed upon between the club and the player prior to January 10 and providing the request for consent to accept such salary is made by the player and received by the
Commissioner prior to January 10."

Twins not hamstrung by this at all.


 

That likely deals with arbitration eligible players that work out contracts between the end of the season and the listed January 10th. Mauer will be a FA next year during that timeframe.

I don't know, is Stewart really much of a risk to get picked? I mean like last year our pitching went from terrible to not quite as terrible, and he was like 9,10,11...on the minor league depth chart. I doubt any team can look at him, potential and all, but look at that and see a place on a 25 man roster. I always thought they should try to work out more trades using the rule 5 though. 

 

And Im all for the wakefield contract, for everyone. $4m/yr mutual option forever? If all the money wouldn't just go to the owners, and they'd make tix and concessions dirt cheap. But if Mauer would be into it I think that would make sense for both him and the Twins.

I find it hard to believe that any major league team would try to hide him in their bullpen for a full season.  His analytics numbers (swing and miss rate, K rate) are just not very good right now.  Can't believe anyone would see enough upside to put him on 25-man all year.

based on how much 1 WAR is worth.

This valuation has been thrown around here quite a bit.  Is one WAR worth $8M or is one WAR the average production of free agents?  Those are two quite different things.  If we attempted to build a team on this premise, the payroll required to build a 98 win team would be roughly $400M.  Our team has to average about $1.5M/WAR.  No team can actually afford a $400M payroll, therefore I don't think one WAR is actually worth $8M, especially to teams with 50 or 60% of the top teams. 

based on how much 1 WAR is worth.

A better measure of market value would be what free agents actually get paid comparted to their average WAR for the previous 2-3 years.  Here are examples of contracts from 2016 & 2017.  I mostly used players in their prime to get a reasonable measure and averaged their WAR over the previous three years to determine market value/WAR..  I did use David Price at age 32 and I omitted the year Cueto had only 60 IP.  I threw Jason Castro in just to add an example of that type of player.

Year | Average War | AAV  |  AAV per WAR

2017 Yoenis Cespedes … 4.4  |  27  |  6.136
2017 Jason Castro … 1.43 | 8.167 | 5.711
2017 Josh Reddick … 2.5 | 13 | 5.200
2016 Jason Heyward … 4.86 | 23 | 4.733
2016 David Price … 5.63 | 31 | 5.506
2016 Johnny Cueto … 4.43 | 21.67 | 4.891
2016 Justin Upton … 3.46 | 16 | 4.624

It is clear that the market does not value free agents anywhere near $8M per WAR.  These players averaged 5.257.  My guess is that RPs might bring this number up a bit.  Cespedes missed some games which brought down his salary/WAR.  If we use 2 previous years instead of three his value is 5.55/WAR which is very consistent with the others.

BTW … Mike Napoli got 7M on a 1 year deal in 2016. That’s about the best comp I can come up with for Mauer.  Totally different player but a similar caliber 1B.

There was a very interesting series of articles by Matt Swartz in fangraphs this summer. He looked at several aspects including varying methods for calculating the dollar amount and the aspect of linearity.

It makes me wonder about Lynn and Cobb. Lynn is projected by steamer to have 1.3 WAR next year. Over 4 years with decline and chance of injury he might project for 3-4 WAR. A generous projection might give him 6 WAR perhaps declining from something starting above 2 next year. Teams might be in at around 50 million for 4 years to buy that hope of 6 WAR. By steamer Cobb projects better next year at 1.7, but a 6 WAR over the next 4 is reasonable though a little generous.

Interesting steamer projected Chatwood at 1.9 fWAR. He is younger and decline isn’t a factor but injury is a factor. If the Cubs projected him at 4.5 WAR over three years his contract is close to market.

A team’s ability to project is incredibly important here. The in house work is critical. Teams are getting better but ERA, wins and holds are still overvalued and overpaid. One of his articles shows a premium for FIP vs. ERA that is still present in the market.

Your SS 7 player sample doesn't make anything clear.

OK Let’s look at ALL of the players who received contracts of 3 years or greater last year.  That would be the definition of assessing market value, right.  I used these players because they obviously were the best positioned of all of the free agents. I also picked the best players because I don’t think most TD participants are much more interested in the Johnny Cueto types as opposed to Jason Castro.  

Yoenis Cespedes … 4.4  |  27  |  6.136
Jason Castro … 1.43 | 8.167 | 5.711
Josh Reddick … 2.5 | 13 | 5.200
Justin Turner … 4.23 | 16 | 3.783
Ian Desmond … 3 | 14 | 4.667
Dexter Fowler … 3.1 | 16.5 | 5.323
Edward … 4 | 20 | 5.000
Mark Trumbo … 1.65 | 12.5 | 7.576
Kendrys Morales … 1.4 | 11 | 7.857
Ivan Nova … 1.4 | 8.666 | 6.190

These 10 top free agents averaged 5.74M per war.  It’s probably an anomaly but Trumbo / Morales & Nova brought that average up.  The other 7 players averaged 5.12M per WAR,
  
RPs  
Bret Cecil … 1.13 | 7.625 | 6.748
Aroldis Chapman … 2.666 | 17.2 | 6.452
Kenley Janzen … 2.333 | 16 | 6.858
Mark Melacon … 1.8 | 15.5 | 8.611
Average … 7.167.

Top BP arms got a premium last year.  I suspect this average comes down if you look at the masses but I am not interested enough to look them all up.

Is the market data from all of the FAs able to get 3 year contracts adequate evidence for you?

I always applaud an analytic approach to decision making.

 

You seem to be aiming toward a forecast of WAR, at the time the contract is signed. I think the studies computing $6M or $8M (or whatever) per WAR are doing it in terms of value delivered, looking back after the contract is signed (so the jury is still out on recent guys). That is, they are asking the question, what do teams typically get for their FA money?

 

I think it stands to reason that front offices' forecasts are getting better and better, so as time goes along the actual value delivered (at least insofar as WAR represents "actual" value) becomes a better and better proxy for the forecasts.

 

But even in trying to come at the question from the forecast WAR for these players, a simple average of previous years seems awfully far from what teams probably use. I don't think you would use that kind of average for a product life-cycle study, for instance, unless at a very particular stage (probably mid-life) - you would draw dangerous conclusions if in the early years, or if at the end in maintenance mode when you're trying to wean customers off. Ballplayers, as a "product", surely aren't static enough in their "life cycles" to try that with.

 

With the horizon you specified, I would weight the most recent year more heavily than the prior year, and much moreso than two years back - almost ancient history in some cases. But I would also factor in growth or (more usually, for these free agents) decline relative to age. Injury risk also comes into play. It may be that teams are (in the privacy of their own processes) putting a very large downward factor on their forecasts over the life of the contract they intend to offer. That would move things in the opposite direction from your conclusion, since it makes the denominator smaller.

 

Another thing that makes the methodology difficult is that players' "accurate" forecast of value, by whatever procedure you think best, is probably in the middle range of what teams will compute for themselves; and it's highly likely that whoever computes the highest value will make the largest offer, and in turn the player is highly likely to accept an offer very near the top of the range. This of course would move things in the direction you suggest, as it makes the denominator larger.

 

All things considered, it's hard to approximate teams' forecasts with a simple average. Beyond hard, I think - misleading, or even not useful. We'd be dividing $$/WAR using something basically unknowable.

 

Cespedes, the first guy on your list, seems like a good example. He was a highly sought Cuban free agent in 2012, and when he became available again in the 2016-17 offseason his resume was a bit spotty. He was coming off a 2.9-WAR season (I'm using b-r.com) after a 6.2, decidedly his best, and he had missed a few games in August after putting in two full seasons the prior years, after starting off with two injury-impacted seasons. I could imagine some widely differing forecasts by competent professionals in the field. When the Mets prepared their eventual winning offer, it's not preposterous to think that another 6-WAR season could be expected. But, he was already turning 31, and for the four-year contract the team was contemplating, a decline could be expected. An injury during any of those four seasons could further harm the value delivered in that season just due to his absence, while also perhaps accelerating his declining ability for future seasons. All in all, just spitballing here, a four-year WAR of (6,5,4,3), times an 80% chance each season for not having a really serious injury, comes out to only about 14 WAR over the life of the contract, or about 3.6 a season, rather than the 4.4 you came up with. That comes out to around $7.5M per. If I did the math right. Again, I'm only spitballing.

 

Now, given that the Mets won this sweepstakes, it's fair to assume that every other team came up with a lower WAR estimate and made commensurately lower offers to the player. Unless you believe the Mets are super geniuses and have a unique ability to forecast future value, it's very likely that the actual value he returns will be lower than they paid for. Ergo, the cost per WAR will likely be higher than this guesstimate. That's the so-called Winner's Curse in any free market, right? (Not that MLB markets are all that free. )

 

I'm not going to invest the time with my rinky-dink eyeballing methodology, on the other players you listed, but you probably see my point, that trying to infer MLB forecasts of WAR is harder than just averaging some recent seasons.

Everything you have pointed out is very fair.  It would have actually lowered the dollars paid per WAR had I used 2 years and I would agree that teams are definitely going to weigh the newest information more heavily.  Had I added some common sense in cases in terms or whether to use the last year, two years or 3 years, the amount per WAR would have come down.  I only used the 3 year approach because it was conservative and others could not suggest I was cherry picking. 

The point I am making is not this granular.  It's this simple.  It is crystal clear that a model paying $8M per WAR is not sustainable for even the largest market teams.  To say its reasonable or sustainable for all of the teams in the league is to ignore many economic and business principals.  I guess it's this simple .... If the MN Twins or any team of similar revenue produce at a rate of one WAR per $8M, the probability of success is extremely low if FAs are relied upon to any significant degree.

That's still way too small of a sample size.
You need to look at all FA's, not just 3 year guys. Also need to look at multiple years.
I also don't know why you are looking at previous to contract WAR. What matters is price per WAR during their contract.

And no team is comprised of nothing but FA's. Guys under team control are much cheaper. That is where you get the value from. Of course $8 million per WAR would not be sustainable for a full 25 man roster. That is why developing talent is so important.

But it is easier to ignore the statisticians and in depth analysis that has been done on a major site like fangraphs.

With that being said - A team should target exceeding the market value when signing FA's. And the other big thing is that teams should try not to pay market value for 1+ WAR players (like Chatwood or Gibson) since you can find those guys on the cheap (like Gibson). I think fangraphs in the past has said that X/WAR doesn't account for the increasing value of elite players since they only occupy one roster spot. Meaning that two 3 WAR players don't equal a single 6 WAR player.

Addendum #2 - I also find the 8M/WAR a little high but it doesn't really change anything if you change the number to 6M/WAR.

I like the idea of an increasing value on WAR

1st WAR - 3m

2nd WAR - 5m

3rd WAR - 7m

4th WAR - 9m

5th WAR - 10m

So... a 2 WAR player would be worth 8m annually and a 4 WAR player would be worth 24m annually

Maybe my escalation numbers are a little low, but you get the idea