Monday, November 30, 2009
One category roar, five category snore
Posted by Derek Ambrosino at 5:06amA 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.
| R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|
| 65 | 13 | 59 | 6 | .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.
| Player | R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|---|
| David Murphy (269) | 61 | 17 | 57 | 9 | .269 | |
| Gordon Beckham (293) | 58 | 14 | 63 | 7 | .270 | |
| Howie Kendrick (243) | 61 | 10 | 61 | 11 | .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.)
| R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|
| 68 | 15 | 63 | 9 | .281 |
The best comparison I found for this line was Melky Cabrera, who was ranked 230th overall.
| R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|
| 66 | 13 | 68 | 10 | .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.
| Player | R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|---|
| 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.
| R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|
| 68 | 15 | 63 | 9 | .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.
| R | HR | RBI | SB | AVG | |
|---|---|---|---|---|---|
| 68 | 15 | 63 | 9 | .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.
Derek Ambrosino aspires to one day, like Dan Quisenberry, find a delivery in his flaw, you can send him questions, comments, or suggestions at digglahhh AT yahoo DOT com.
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.