Mock Battle: Ryan Braun vs Michael Brantley
You can follow along as we mock, but you’ll see us highlighting specific picks and rounds over the next couple of weeks too. Multiple times in the draft, I found that there were back-to-back outfield picks that were interesting. Matt Kemp and George Springer. Michael Brantley and Justin Upton. Yasiel Puig and Carlos Gomez.
In the third round, after picking up Miguel Cabrera and Anthony Rizzo with my first two picks, I found myself in just such a predicament. I wanted an outfielder, and Michael Brantley and Ryan Braun were clearly the best of the remaining. Given their respective ages, histories, name values, and draft positions, I found that juxtaposition interesting and worth investigating.
Would you rather have Brantley or Braun?
Full disclosure. I have not yet done my own rankings, since I’m deep into editing FanGraphs Plus, so I’m winging this mock. Once I investigate, I may decide I made the wrong decision. So it goes.
The first thing to do is check the projections, of course.
| HR | R | RBI | SB | BA | |
|---|---|---|---|---|---|
| Ryan Braun | 24 | 79 | 80 | 13 | 0.278 |
| Michael Brantley | 13 | 75 | 72 | 14 | 0.290 |
By Steamer projections at least, Braun gets the advantage. Per Jeff Zimmerman’s Standing Points Gained research, homers and steals are roughly equal in terms of owned stats in actual fantasy leagues. Since Brantley (637) and Braun (610) are roughly equal in their plate appearance projections, every homer is worth about .0033 in batting average. That means Braun’s ten-point lead in homers plus stolen bases overcomes his 12 point deficit in batting average.
But, contrary to some public opinion, I do not run my drafts on projections alone. And if you’ve listened to our (FSWA-nominated) podcast, you’ll know I’m no dollar value fascist.
A quick synopsis. If you choose a projection system, if you make projections yourself, if you manipulate projections to produce your own set — these are subjective changes. You’re guesstimating. We all are, and projections themselves only explain about half the variance in outcomes in any given year. To then turn around and yell about ‘not spending one dollar more’ than you’ve got on your spreadsheet makes no sense. Especially once you then introduce the randomness that is an auction / draft room / actual results. Suddenly you’re interacting with other human beings that are also making subjective decisions. So yeah, sometimes I’ll go a dollar or a few over my values for a guy I like. I could lie and tell you my values say to spend what I spend, but instead I’ll point to the impossibility of reducing the entire auction experience to one number. Yes, I wrote that on FanGraphs.
Back to the lecture at hand. A few aspects of the projections for these two players feel particular shaky.
Ryan Braun’s playing time, for one. 610 plate appearances would be a three-year high, and he’s 31 years old now. Between a nerve injury in his right hand and a few oblique strains, he’s missed some time. But once again, using Rob Arthur’s injury projection for position players, we find that these obstacles only predict eight missed games from Braun in 2015. Not even a full DL stint.
Ryan Braun’s power, too. After those years of near-.300 isolated slugging percentages, the .203 projected ISO seems modest. But really it’s right in line with the most likely non-steroid outcome that I outlined before last season. Over 600 PAs, it looked like he’d hit 24 homers based on evidence that steroids can lead to bat speed through muscle mass. He would have hit 20 last year with 600 PAs. Given that he played through an injury that has since been corrected by surgery, and Jeff Zimmerman showed that playing through injury can suppress power projections, there’s a tiny bit of upside remaining even if Braun is off the juice.
Michael Brantley’s power, of course. I took a much longer look at this already, and though it’s nice that his power came from an identifiable source (pulling the ball), the fact that his batted ball distance didn’t change much is a warning sign for the soon-to-be 28-year-old. His overall game is enticing, he’s a good real-life asset, but I’ll buy Steamer’s projected power.
I still feel pretty good about this decision, but I also think Michael Brantley went at right about the right place. If he retains more of his power than the projections believe, he’ll also likely hit for a better batting average, and if both of those things happen, there’s still a good chance that Brantley will out-earn Braun in 2015.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
Ellsbury for 29 or Heyward for 13 with 40% inflation? (roto with OPS as 6th category)?
as an Ells lover with mild disbelief in Heyward, you gotta go Jason here. the price difference is pretty staggering! sure, Ells has the better 2015 given health, but at more than twice the price? go Heyward.
…age and the inclusion of OPS only strengthen the case
I think Brantley’s counting stats are very conservative via Steamer. I see him being mid-80’s in both R and RBI. His contact rate is also very reassuring. The HR regression is happening, but even with that I take Brantley all day long over Braun. Too much uncertainty with Braun to warrant paying a high price.
I almost posted the same comment citing a healthy Kipnis and Moss as reasons to be more optomistic about his R+RBI, even going as far as saying that 180 wasn’t an unreasonable total if he contimues to bat 3rd in the lineup.
I left R+RBI alone because I felt both were light. Brewers have a decent offense too, and if Braun really hits that PA threshold, I think he’ll do better than his R+RBI projection.
I’m not as down on the power, though of course it will likely regress, in terms of batted ball distance; while having essentially the same FB-IFFB% the last few years, his FB distance over the last four went 266-269-276-280. That’s a fairly straight-up trend to me. Also don’t see his SB drop as much as projected, for while a near-100% success rate can’t continue, he did show much improvement on that in 2013, as well.
you guys should use jeff’s sgp work in your dollar values. More useful than zach’s z-score method in my opinion. thanks for linking to that.
More relevant than the steroid thing (IMO) is his hand injury. He finally had surgery on it this off-season, but it’s the kind of injury that has the possibility of staying with him for the rest of his career.
Whether it succeeds long term is indeed the question, but I’m not sure it could properly be termed surgery. I don’t think they opened up his hand. He got “cryo” treatment on the nerve endings.
I’d be a little leery about Braun’s thumb surgery. I believe it had not been performed on a thumb before? I’d probably wait for spring training to start before taking that plunge.
Cueto. I like this pick. Where are you with his injury concerns?
“To then turn around and yell about ‘not spending one dollar more’ than you’ve got on your spreadsheet makes no sense. ”
your move larry..
If you have to go a dollar or two more on guys you like, then you’re simply doing a bad job in the projection/valuation process. Was the projection wrong, or the valuation process itself? Who does that dollar get reallocated from? Is it a systemic issue that you are only correcting in this one instance and you don’t notice all the small errors throughout the rest of the draft/auction? People toss away a dollar here, a dollar there, and it adds up. $5 of wasted auction budget is the equivalent of trading your 3rd round pick for a 5th round pick. For casual fantasy players I can see the argument of diminishing returns. But when people spend this much time on fantasy baseball analysis, I think it’s fair to expect it to be more thorough.
I’d say that’s a false sense of security. Even if you go through and change projections by hand, you’re making subjective decisions that affect your dollar valuation at the end. I’m not saying I don’t use my dollar values derived from projection systems, but in lieu of patting myself on the back for moving stats around on the player level instead of on the dollar level, I’m willing to spend on the players I like.
Let me put it this way. There’s a run environment in projection systems. Every run scored and allowed is accounted for. if you’re doing hand projections and change a pitcher’s ERA to a lower number, are you taking those runs from one player and adding them on somewhere else?
If not, you’re just putting the dollar on the player on the stats level.
great point
No, I don’t believe that the run environment needs to even out within a projection set. There’s no reason for that to be the case. Overall MLB-wide, obviously the run environment is zero sum, but not within the subset of players that are fantasy relevant on draft day. Also, we know that a lot of pitchers will be injured and replaced with significantly worse pitchers, as with hitters. The pool of crappy replacement players that will never see time on a fantasy roster is a buffer for the rest of the league. The best we can do is approximate the value of a player knowing a presumed run environment which includes all of those replacement pitchers. Masterball used to deal with this when they tried to strictly stick to a perfect run environment, but they have rightly changed their tune in recent years.
Whereas adding a few dollars arbitrarily to players you like does indeed affect the value of other fantasy relevant players, especially if like I said it was a systemic issue that lead to the poor initial valuation (ie: all SB threats are over/undervalued, use of inaccurate hitter/pitcher split, not accounting for injury risk, etc). The difference here is that we can easily see who those extra dollars should be given to/taken away from. Sure you can adjust Brantley on the fly by $2, but what about the other 20 or 30 guys that you undervalued by $2 but may not have bid on, and on the other end we now have a whole lot of guys that are overvalued by $2 to balance it out, who happen to be looking pretty tempting since they’re going a couple bucks under our value for them. If you’re off by a couple dollars on a significant amount of players, those are the ones your team is going to be full of at the end of the auction, and the ones you undervalued get taken by other owners for maybe a 1 or 2 dollar discount. Systemic errors can add up pretty quickly. If you like Brantley more than Braun or vice versa, that should be reflected in the valuation. Identify what it is that makes you like them and give it a value.
It’s not a false sense of security to trust your values either, all else being equal. Team needs, auction inflation, among other things can change values (and you can change all values to reflect it in real time), but there’s no need to start making on the fly adjustments to random players in the third round of a straight draft. The error within those projections gets magnified by the inherent variation in outcomes. Just like betting on a coin coming up heads two times out of three flips on account of the fact you know the outcome won’t be exactly split 1.5 each, and you have a good feeling about heads.
These are all good points, and why I normally do stick to my values. But I’m not going to be crazy about it, there are always two or three guys where I go over to get them.
I’m still not convinced that going into a player’s projection and just adding a few RBI here or taking few runs away there is all that scientific. I’ve always been a feel guy. A feel guy with a good set of values that I trust, but a feel guy.
And projections are only 70% accurate so there should be a touch of wiggle room on the dollar values, too.
Worse than that. The year to year correlation on projections is .7, but that means it only explains just about 50% of the variance in the actual outcomes.
Right, there’s variance. So the further you stray from the best possible projected mean, the bigger problem variance becomes. There’s also a lot of variance in outcomes when playing Blackjack but that’s not justification to use sub-optimal strategy.
Skin, I want you to place a 95% confidence interval on your projected performance and then convert it into a dollar amount confidence interval. Then let me know if it’s ridiculous to overpay by a couple of bucks here and there. Projection systems are far from omnipotent and either you have entered your personal biases when ranking by making subjective changes to the projection system, or you use an unbiased projection systems and then account for factors after the fact that are not incorporated into the projections. Take masahiro tanaka for instance. You know he had a partially torn UCL midway through last season and rehab and time off ameliorated the partial tear. So your projection system says 40% UCL tear 60% no tear and he will be just fine. Using a priori uniform distribution for when the injury might occur, we would expect him to miss half the season 40% of the time therefore he is projected for 160IP (using a rough proxy of 200IP per full season). Now is this the most likely outcome? No because he has a 60% chance of ending the season unscathed which would be the maximum likelihood estimate. Now you take your projection and say oh well I’ll pay x dollars for this projection and stand pat on it? This is without accounting for risk aversion, weighting the playoffs and regular season in an appropriate manner (assuming your league has a playoff system) and taking into account of performance variation. So then say you’re confident about your values and you keep on getting outbid on the first couple of position players and then are able to buy star pitchers under your sheets, or theoretical values, at what point do you say, okay maybe my market prices are off and I should adjust accordingly so that I don’t end up with a star staff and scrubs lineup? Because if your saying the third round is too early I’m afraid you would wait too late to adjust to a dynamic draft.
That being said, everyone has there own style and chances are it works for them, otherwise they probably wouldn’t be reading roto graphs. I just side with eno et al on this one.
That whole Tanaka thought experiment doesn’t apply to what I’m arguing… my point is you make your values (using whatever method) and then stick to them unless something in the draft changes how you should value them. The fact that you like Ryan Braun more than Michael Brantley isn’t a revelation that spontaneously arose during round 3 of a draft. You liked him when you made the values, so you should assign a proper value to him. Maybe for somebody that uses pre-made values, they may want to adjust mid-auction for the guys they like more than the projection system suggests. Although this would be for somebody that is only interested in doing 10 minutes of draft prep. It’s better than doing nothing at all, sure, but it’s by no means a good idea for somebody who is reading/writing fantasy baseball articles in January.
Overall, is it a big deal? Not really, unless like I said it’s a systemic issue, in which case you’ll wind up with a group of players that you systemically overvalued. But it’s hardly a badge of honour to say “I don’t trust my own values”. And it’s not so much being dollar value fascist to stick to your numbers. It’s about being prepared. If you decide at the last minute “you know what, Braun’s injury isn’t that serious and I don’t know if Brantley’s HR/FB is sustainable, I’m gonna bump Braun up a notch” then you are being just as much a fascist for values, you’re just not deciding on what that value is until the very last second.
I’d love to draft with Skin; drafts are fluid, and you have to adjust based on what others are taking or else you’ll find yourself with a $2 guy at some position who is a “value” because he should be $4 but, well, I’d rather have the $20 guy taking up a valuable active slot even if I pay $24 because others have been “overpaying” for that particular position.
You must not have read the part where I said the following:
“Team needs, auction inflation, among other things can change values (and you can change all values to reflect it in real time), but there’s no need to start making on the fly adjustments to random players in the third round of a straight draft.”
The adjustment to Brantley’s value in the third round is a gut feel change of heart, not a reaction to a quirk in the draft.
Hard to bank on the power from either guy at this point. Braun simply stopped pulling the baseball last year. Is his bat speed going to come back? who knows? I dont. But his lack of pull power last yr doesnt sit well with me. I wouldnt count on either hitter reaching 20hrs, but I do think Brantley has a much greater chance at sustaining sbs and avg. I wouldnt be surprised if Brauns sbs dipped to single digits. He is projected for 13 after stealing 15 over his last 830+PAs. I think the under 13 on Brauns sbs is a safe bet. Give me Brantley because of what I believe to be a higher floor due to safety in avg and sbs.
I remember the days of getting The Sporting News Fantasy Baseball magazine and looking at the dollar values. When I went to my first auction, I tried to trust those dollar values. (I know: How naive.) I was totally ready to spend (gasp!) $40 to get Barry Bonds or Albert Belle or Ken Griffey. I ended up with zero superstars and a pile of guys that cost in the $15 – $18 range, and a bunch more in the $5 – $7 range. I got absolutely smoked. That price list was not just useless, it was a recipe for finishing at the bottom of the standings.
For me, the takeaway was you can prep all you want but the reality is that auctions are unpredictable. There is more inflation some years and less in others. Your dollar value on Brantley might be $25, and you might love the guy and be ready to pay that $25, but if the bidding goes to $26, I don’t think it is any kind of crime to go to $27. My experience is that being married to a set of dollar values before the auction is not what leads to winning.
To me, the takeaway from that example is that you should do all the draft prep that you can, rather than blindly using values out of a magazine. Coming to an auction prepared with realistic dollar values for players doesn’t mean you’re married to them. But if you think Brantley is a $27 player then why wasn’t he valued that way prior to the auction?
Okay I guess I slightly misunderstood. It seems you are more suggesting that within a position group a manager should have his ordinal rankings before going to the draft. I guess the point I was trying to make is that if you slightly prefer Braun to Brantley so their respective values are $26 and $25 and Braun gets bought for $30, I find it hard to refuse to make a $26 or $27 bid for Brantley if you feel the marginal upgrade he is over the next CF is worth it.
The other thing I was thinking is that a $26 Hanley Ramirez or $26 Bumgarner shouldn’t be absolutely preferred by the manager over a $25 Brantley by the third round as plenty of surprises happen that quickly. Just look at how polarizing some of these picks in the mock are by experts and then amplify it by the prejudices of 16 individual managers (or how many teams you have).
Finally, if differentiation is not applicable then I guess I fail to see the distinction between “Not spending another dollar or two because it adds up” and “Prepared with a realistic dollar values for players doesn’t mean you’re married to them”, and also the difference between “Team needs, auction inflation, among other things can change values (and you can change all values to reflect it in real time) and “but there’s no need to start making on the fly adjustments to random players in the third round of a straight draft” to me in real time and on the fly are nearly synonymous. Once you create your rankings I assume you change them to account for new information as you get closer to draft day after you’ve “completed” your rankings. Maybe spring training numbers, other research articles, or articles by beat writers detailing the shape or offseason preparation should, if you find it compelling, change your opinion of a player. In which case your ranking should change every time with the lone exception if you have 100% a priori belief in your rankings and none in the posterior observations. Adjusting for these subjective changes can be challenging when trying to create an objective projection, so making any changes away from our original projections is something we don’t forget. Because of the “inherent variations in outcomes” thinking you over or under adjusted on draft day is a consideration most managers who tinker with their projections and rankings have to assess in the war room. Thus any evidence on draft day that you might have gone the wrong way, you should be able to adjust accordingly. Just my two cents.
This writer knows nothing. Over the past 3 years Brantley is up in almost every category and way ahead in some. Brantley is by far the better player and is still getting better, while Bourn is getting worse.
He knows the difference between Braun and Bourn.