Thursday, March 12, 2009
Tiering up
Posted by Jonathan Halket at 1:28amA hot piece of advice these days is to focus on ranking and drafting players using a tiering system. The reasoning behind the system is that the best talent (in the tail of the distribution) is thinly distributed—the difference in expected value between the top two players is usually larger than the difference between the 99th and 100th players. Moreover, it is sometimes possible to see a noticeable step down in value between two consecutively ranked players; for example, there could be a larger than normal gap in value between the third and fourth ranked players at a position.
The most cited example of this is at shortstop, with Hanley Ramirez, Jose Reyes and Jimmy Rollins significantly better than Alexei Ramirez, Stephen Drew and Troy Tulowitzki.
A tiering system has two aspects: a ranking/forecasting part and a drafting strategy.
The ranking part of the tier system recommends paying more time and attention to slotting players into tiers, both overall and by position, rather than worrying too much about rankings within tiers (i.e. getting more exact dollar values or rankings). In general, I think this is useful. I would rather have system that put all of the players into their actual, true tiers but then got the ordering within the tiers somewhat wrong, than a system that got many of the exact rankings rights but messed up on some players' tiering. In other words, many small mistakes are usually better than several big ones.
This comes with an important caveat though: the exact location of tiers gets harder to find the further down you go. The deeper the position, the easier it will be to find a third and perhaps fourth tiers (like at starting pitcher or outfielder). Still, meaningful differences between merely tiering and exact slotting get harder to notice as you go deeper into the ranks. It is pretty easy to get tied up in logical knots. I fooled around with tiering the outfielders and I quickly ended up with a quandary: I don't think Nick Markakis is equal to Manny Ramirez, but it is equally hard to find a way to divide the outfielders ranked between them into two distinct tiers.
The drafting in tiers strategy basically boils down to the following type of advice: it is your turn in a draft and you're trying to choose between Rollins and Chase Utley. Hanley and Reyes have already been drafted, so Rollins is the last of his tier left while Ian Kinsler, Dustin Pedroia and Brandon Phillips are all still on the board, but you have Utley ranked higher than Rollins. Whom should you draft?
The tiering strategy, if you're a purist, says that you should draft from the bottom of the tiers whenever possible. In this case, that would be Rollins over Utley. Of course, it is rarely this easy.
A much better tiering strategy should also take other things into account. For instance, is someone from the alternate tier (like Phillips) likely going to be still available when you next draft? How much worse is the difference between Rollins and Alexei Ramirez than the difference between Utley and Phillips and between Utley and Brian Roberts (assuming he's the best of the next tier)? How far is drop to the next respective tiers?
If you're drafting at the tail ends (picks one and 12 in a 12-team league) the tiers have to have rather more players in them for tiering to be useful. If you're not going to pick again for 22 picks, what's the likelihood that any players in a three-player tier are left? And if you have to start worrying about the relative differences between players, then you're going to need a ranking system that makes finer distinctions between players than mere tiering would do, which kind of obviates the whole point of tiering in the first place.
If you have a question for the Roster Doctor email here. Emails in simple text with players' full names properly spelled are much more likely to get responses. Also be sure to include your league's player pool (mixed, AL-only, NL-only), number of teams, scoring format (roto, head-to-head, points, etc.), categories, whether or not it's a keeper league, and any other pertinent information.
I’m not a fan of tiering, since I think it puts an artificial value bounds on players that exist more in a continuum.
In your example, you take Rollins over Utley because Rollins is the “last of the first tier” while there are many others at Utley’s “level” available.
I think that’s the right decision, but I think this is a false semantic distinction and you don’t need to invent “tiers” to make it correctly.
Rollins is the right choice because at the moment of your pick, there is a scarcity of high value at shortstop, while there is a wealth at second base. You don’t need tiering to make this decision, just basic economics.
Think of it as position scarcity (which most understand and use to some degree) that changes on a pick per pick basis. A draft is a rapidly evolving market, in which the supply-side of various items (shortstops worth having, second basemen worth having) goes up and down from turn to turn.
Rollins might be worth more than Utley in some drafts, at some turns… but in other drafts that develop differently, Utley would be the right choice. This has much more to do with the dynamic supply than any tiering abstraction.