One category roar, five category snore

A couple of weeks ago, in the comment section of my Approaching unconscious competence article, a discussion sprang up regarding the way a player’s value against replacement is dispersed across categories and how it affects the perceived and real value of that player and his tradability. I touted the trade value of one-trick ponies while reader, Andrew P. advocated the all-around marginally better-than-average-type of player as being easier to trade.

I decided to try to explore this in a very unscientific way, basically for food for thought purposes. This is not going to be a fully in-depth study with peer-review level mathematical rigor. The most honest reason why I’m omitting the heaviest mathematical lifting is that I’m not great at it, and doing so might just devolve the discussion into an argument over methodology. But, also important is the fact that when discussing things like trade-ability we are talking about perception. And, since you are likely not in a fantasy league with 11 Tom Tangos, the way other owners perceive value of [your] players is important, and not necessarily congruent with their actual values.

My argument is that when it comes to players with overall marginal value (i.e. last roster spot, bench depth, trade bait, etc.) the player whose value is highly concentrated in fewer categories is the more attractive commodity.

Before we get into looking at some players and where their value comes from, let me make two semi-tangential points. First, when it comes to player production throughout the season, studies show that streakier players, that is players whose positive value is packed tighter into a smaller bunch of games, are actually worth more wins over time than those who display consistent, but non-dominating success. This is logical when you think about it, as a player who goes 4-for-4 in a game will likely influence that game much more than his composite influence on four consecutive games in which he goes 1-for-4 in each.

Second, if any of you out there play fantasy basketball, you are probably familiar with the legend of Zach Randolph (and, to a lesser extent, Tony Parker). Every year, one of two things happens to Randolph. Either some naïve owner sees the 20-10 and drafts him a good 40 picks too early, or some astute owner drafts him, and then packages him with a top-40 player and trades the two for a naïve owner’s first-round stud. Randolph is a liability of epic proportions in most categories, but puts up such attractive numbers in the glamor categories, that his perceived value is light years beyond his actual value. Ditto Parker, by the way, a 20-8 point guard who is not worth a top-70 pick.

In different ways, both of these points are illustrative of both the real and perceived value of tightly concentrated production.

With that in mind, let’s get to some fantasy baseball. To start this investigation, I wanted to establish the replacement level in all the five standard categories. I took a standard 12-team mixed league with expanded rosters, totaling 12 starting bats per team. At 12 teams, and 12 roster spots, that makes 144 starting offensive roster spots. So, I simply went down the line and cut off the 144th most productive season in each of the counting stats. For, batting average, I just took the average of the team batting averages that earned six points in all 12-team mixed leagues I played. So, let’s take a look at what a player’s line would be had he produced at the 144-cut-off in all categories.

RHRRBISBAVG
6513596.279

Now, it was difficult to find players who fit this model, but here are some players who were roughly similar across the board, with no large chunk of their value coming from a single category. Their overall Yahoo ranks are in parenthesis.

PlayerRHRRBISBAVG
David Murphy (269)6117579.269
Gordon Beckham (293)5814637.270
Howie Kendrick (243)61106111.291

Now, let’s take a look at what the marginal five-category contributor might look like. My first inclination was to just bump up each category by 10 percent. But I nixed that idea because runs and stolen bases, for example, are economies of two different scales, and proportions would be misleading. (See, I’m not totally mathematically naïve/illiterate). So, instead, I chose to redo the cut-off to determine what the 10th percentile of the roster-able players looks like, meaning the cut-off now comes at 130 as opposed to 144. So, here’s what the five-category 10 percent above-replacement player looks like. (I just moved up the batting average to the average of the “7”s in my 12-team leagues.)

RHRRBISBAVG
6815639.281

The best comparison I found for this line was Melky Cabrera, who was ranked 230th overall.

RHRRBISBAVG
66136810.274

The first thing that we should notice here is that none of these players played entire seasons. Cabrera logged 485 ABs, while Murphy logged the most ABs of the player on the replacement-level list, with 432. Melky Cabrera was no fantasy world beater, and apparently being slightly better than 12-team fantasy replacement value doesn’t earn you a full season of ABs on the New York Yankees either.

Let’s take a quick break for pro-active disclaimer/friendly reminder. Now, of course Kendrick and Beckham helped fantasy teams because they are both very talented players with high ceilings who produced their value in less than full seasons, and their owners got production from other players in their absences. The seeming irony is that both of these players were more helpful than Cabrera (because of that concentrated production). In reality, we know that when it comes to the last few roster spots, good owners cobble together composite production from a variety of players. But, again, for our purposes here we are just looking at where individual production cut-offs are at certain levels against replacement. So, let’s try to keep in mind that we are looking at seasonal production for the players in question with minimal concern for how many games or ABs it took them to put up those numbers (the obvious exception being batting average, because the “weight” of individuals’ rate stats is important in terms of determining your team’s whole).

Back to the issue at hand, is Cabrera a tradable commodity in a 12-team, mixed league? I don’t think so. Did anybody out there who owned Melky in a similar league receive one-for-one trade offers involving Melky? If you owned him, is there anybody who would have been able to help your team that you think you could have gotten for Cabrera? When it comes to trade returns, to paraphrase John Sterling, the Melk Man doesn’t deliver. If I wanted to trade Melky Cabrera in a 12-team league, here are the situations in which I would think it might be possible.

I might be able to land either an elite set-up man, or a set-up man behind a troubled closer and get some value out of it. If I was Ryan Madson’s owner and was having health issues with my outfield, I could see making a Madson-for-Melky deal. But, more likely, I’d try to take a couple of swings for the fences on the waiver wire.

I could maybe see the Melk Man involved in a two-for-one. If, for example, I was trading a high quality all-around OF, say Nick Markakis, for high level starter or closer, I’d consider asking for Melky as well, as a throw-in, if he would be a better replacement for Markakis than anybody on the wire. Of course, I have to be willing to drop another player (most likely a pitcher) on my roster to make room for Melky. (There really are no such things as two-for-ones, because of roster constraints, which is what makes it difficult to add a marginal player in a two-for-one deal, as the owner receiving that player still has to judge that player against his own marginal players.)

So, Melky Cabrera, who is all-around above replacement value is pretty much worthless on the open market in a 12-team league. This is because of the motivations behind trades. Teams make trades, usually from a surplus, to make up ground in categories. A player who 10 percent better than replacement across the board will have no profound impact on the standings in any individual category, and is therefore unappealing as a potential acquisition by a team looking to trade.

Now, let’s look at a few players whose value in most categories was pretty close to the replacement margin, but who produced particularly well in a single category. Overall rank in parenthesis again, as are deviation for the replacement level in each category.

PlayerRHRRBISBAVG
Marco Scutaro (112)100 (+61%)12 (-8%)60 (+2%)14 (+8) SB.282 (+.003)
Adam Kennedy (166)65 (+5%)11(-15%)63 (+7%)20 (+14 SB).289 (+.012)
>

Kennedy’s production is virtually even in the R, HR and RBI departments, taking into consideration overs and unders, while sporting considerable advantages in batting average and steals. Both of those categories are clear, actual assets that Kennedy offers.

Scutaro’s batting average advantage is negligible, and while he does provide speed that will contribute to your team, his big difference comes from the fact that he is actually elite at one category (runs), even though he is virtually a replacement level in three of them (HR, RBI, AVG.)

Once again, let’s look at our fictitious +10 percent all around player.

RHRRBISBAVG
6815639.281

I know that the absolute values versus the relative rank dynamics of the stolen base category throws things off a bit, and that Scutaro especially has a bit more value than our hypothetical player regardless of how it is distributed, but don’t both Scutaro and Kennedy’s lines look more trade-able?

Before I try to summarize some of the takeaways of this somewhat unfocused and non-scientific column, let’s take a look at one more player, Franklin Gutierrez. Gutierrez had a very good year and turned out to be a surprising source of five-category contribution. Without putting up staggering numbers in any one category, Gutierrez was clearly a valuable asset to fantasy teams last season, and he did so in one of the most under-the-radar ways possible.

RHRRBISBAVG
6815639.283
+37%+38%+19%+10 SB +.006

Gutierrez didn’t reach any notable categorical milestones. While he achieved 34 combined homers and steals, he didn’t accumulate as many as 20 of either. He didn’t score 100 runs, or drive in 80 runs. But he did provide meaningful value above replacement in all other counting categories. I don’t think people realize how much marginal value players like Gutierrez provide across the board.

I say this by no means to attack Andrew P., but I think that when he argued for the plus-10 percent across-the-board player, he was thinking of a player like Gutierrez, who was more like plus-30 percent across multiple categories. That’s why five-category studs are so few and far between; they have to be more than just a little better than average at everything to really make their impact. They have to be a lot better at several things. A player who puts up a line of 90-20-90-20 is basically providing approximately 150 percent of replacement value in four categories.

Let’s try to make some sense of what we’ve talked about here and draw some conclusions.

  • When it comes to trades, the way others perceive the value of your players is even more important than how objectively valuable they actually are. And the way that value is distributed across categories affects the perception.
  • Other owners target players whose skill sets are likely to affect the standings. This consideration is, to varying degrees, disassociated from the overall marginal value a player offers. In some cases, this is rational, and in some cases it isn’t. A player’s value to my team may be different than it is to your team, because we are not competing against replacement level; we are competing against each other in team-wide category totals.
  • Milestones are arbitrary and elevate the price of a player more than they actually reflect his value. If Franklin Gutierrez went 20-20 this past year, his ADP would rise some obscene number of spots this year. His 18-16 is certainly helpful, and you can likely grab him for peanuts again next year.
  • I am no mathematical savant.

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: Book excerpt: “Evaluating Baseball’s Managers” – Charles Comiskey
Next: Baby he was Bourn to run »

Comments

  1. Andrew P said...

    Don’t worry, I don’t feel attacked in the least.  You’ve become one of my favorite fantasy writers for your analysis, and I welcome the discussion, especially when there are diverging opinions.

    A large piece of my argument, though, was that someone who produces across the board allows you more flexibility to trade other players on your team without the corresponding hit to one category.

    If you’ve got a one-category player, odds are you acquired them to fill a need in that category.  By trading them away, you again risk becoming deficient in that category.  Even if a more well-rounded player has less perceived value among your leaguemates, you still have the flexibility to trade other more highly-regarded players without the expense of taking a big hit to a specific category. 

    Your production is more diversified and thus, in my opinion, more flexible.

    In the original article, you posted a nice counterargument, that by taking a one-category player, you’d artificially control more of the supply of a certain category, thus driving their trade value up.

    It was a very good point, and depending on the context, could very well nullify my argument.

    As always, I appreciated this article and look forward to more like it in the future.

  2. Derek Ambrosino said...

    Thanks a lot, Andrew.

    Your point about making it easier to trade another player is a good one, but probably moreso in the theoretical than in the actual. When we are talking about overall marginal type players the difference in the non-specialty categories may even be too slight to perceive. Is a difference of a few runs of RBI even something you’d give credence to when making a trade midway through the season. For example, if you’re thinking about making a trade at mid season and note that one of your players has 8 steals while the other has two, meanwhile the first player only has 32 runs, while the other guy has 38, are you even confident pegging the second player having a real advantage in scoring runs, meaning you deem it likely that the gap will repeat in the second half of the season? I don’t know. Maybe, I do, maybe I don’t…

    As far as manipulating the market, I think this is totally possible with saves and steals, where sometimes the second-hand market can be more lucrative than the retail market. This is because while during the draft, steals/saves are but one category of many that you must compete in, after the draft the represent the categories in which there is most to gain by acquiring a single player. I think it’s actually rational, and sometimes even savvy, to draft these categories with the intent to trade them later. The other advantage of stockpiling those categories is that you can control where that production goes. You can either trade them for your own needs, or trade them to help others pass your competition in those categories. This is a much more viable strategy with categories where the number of players who produce them are limited.

  3. Dennis said...

    IMO, these are just two different routes to success.  I am like Andrew.  I aim to build an incredibly well balanced team that is largely invulnerable to all except the most extreme injuries.  I may fall short by a point or two in a category where my team collectively underperforms, but I often make it up in another one, where my players collective overperform.  Because i’m not dependent on the 60 SB guy, if one player contributing to my SBs gets injured, I can find someone that makes up a significant share of the loss, without sacrificing huge points in another category.  In short, I like a buy and hold strategy, trusting that my initial assessment of players and what I need to win a category wins out.  Not surprisingly, I rarely trade, and often have trouble finding trades for the exact reasons Derek says.

    The strategy Derek advocates is a “day trader” strategy, looking to exploit market inefficiencies and put himself in a position to benefit.  He even tries to create them, by using the draft and other moves to create artificial shortages and surpluses.  While it involves some risks in not guessing right on where and how inefficiencies will appear, it can certainly work as well as a buy and hold strategy.

  4. Cody said...

    Derek, I think the argument you frame in the first few paragraphs is a very interesting one.  But I don’t think you can make your point by comparing Marco Scutaro (72.6% ESPN ownes) to Melky Cabrera (17.0% owned) and David Murphy (4.5%).

    By your own admission, Scutaro and Kennedy were roughly replacement level in 3 categories and significantly above replacement value in 2 categories.  OF COURSE they are going to compare favorably when the comps you propose are replacement level in all five categories.  This is reflected by the much higher Y! rankings for Scuatro and Kennedy than for Murphy or Cabrera.

    Scutaro and Kennedy (17.9% owned) don’t really fit the one-category mold you describe.  They’re more like Well Rounded+.  They do two things well, but everything acceptably.

    If you really want to have the discussion about perception and trade marketability for players who are marginally rosterable, the one- or two-category studs you use as examples need to be appreciably below replacement value in the other categories.  Guys who do one or two things well, but everything else poorly.

    I think these would have been fairer examples for your side of the argument:
    2008 Michael Bourn: 57-5-29-41-.229
    2009 Dexter Fowler: 73-4-34-27-.266 (13.7% owned)
    2009 Chris Coghlan: 84-9-47-8-.321 (33.6% owned)
    2009 Erick Aybar: 70-5-48-14-.312 (13.5% owned)
    2009 Skip Schumaker: 85-4-35-2-.303 (11.7% owned)

    To be clear, I think any one of those guys is still a more appealing trade target than Cabrera or Murphy, for an owner who needs a boost in those categories.  So I agree with your conclusions about perception and tradeability.  In fact, I think by using truly marginal players as comps, your point gets that much stronger.

  5. Derek Ambrosino said...

    Dennis:

    Actually, when it comes to the core of my team, I am much more conservative. I try to build a team so that I am not reliant upon my marginal players for large chunks of the production I need in order to be competitive in any single category. (Any player who steals 60 bases is, by definition, not marginal) My post is really more about the question of how you fill out your roster with your later picks, and what the (trade) value is of a player who is neutral against replacement in 4 cats, and 50% better than replacement in the other one, versus a player who is 10% better than replacement across the board.

    Sometimes, players who are overall marginal (or worse) have fantasy assets beyond their real skills, and owners then have to gamble on risky “real” players to produce highly valuable fantasy production. That’s a slightly different situation, and it happened to me this year in my main league. I neglected speed for too long in my draft because there were other players on the board I felt I couldn’t pass up. I was getting 15 here and 10 there, but was developing four-category monster OF, lacking steals. I decided that instead of reaching a little bit with every pick, I would wait a while longer and reach a few rounds on Willy Taveras (I figured his OBP wouldn’t cost him a job with the oblivious Dusty Baker at the helm). Needless to say, I was chasing steals a bit all year. Also needless to say, I would have been totally fine if I had pegged Michael Bourn instead. However, the fact that I tried to stockpile some cheap, possibly trade-able, speed helped me stay competitive enough to finish middle of the road in that category.

    Cody:

    I thought about including Aybar in the comparisons, but the .312 BA just couldn’t be considered marginal. Fowler was a totally different animal, as he was highly above replacement in one category, 41 steals is at least borderline elite, and then below average in three of the five categories.

    In using Scutaro I was just trying to make the point of where a players value over replacement comes from. If we ignore steals (because %s are misleading with small scales), Scutaro is +55% above replacement in R/HR/RBI combined (+61/-8/+2). Let’s forget Melky for a moment and take a theoretical replacement, let’s put him at +18 in each of those categories above replacement, that would make him about 78/15/70, how does that compare in perception with Scutaro’s 100/12/60?

    The fact the cat in which Scutaro has the advantage is the one with the highest scale helps (or hurts) too, in terms of perception. That is actually one of the underlying points of this whole discussion. It takes a little bit of mathematical acumen to adapt differences in raw totals across categories for scale. A difference of six steals is more important than a difference a 12 runs. This is one of the things to keep in mind when trying to determine how others may value your players.

    BTW, I had just chose Melky to try as a decent representation of the +10% vs. replacement in each category. No real player actually fit that mold perfectly.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current ye@r *