Thursday, February 11, 2010
Falling backwardPosted by Jonathan Halket at 6:10am
When it comes to draft or auction strategy, it is always extremely useful to plan ahead and to think about what you want to do later in the draft/auction (from now on, I'll just refer to a draft) when planning your first moves. But, don't get too caught up in your masterful late-round strategies.
Whether you're the type of player that comes into a draft with a full ranking and/or dollar values for all players or if you're the type that comes in with a few gut feelings and casual sentiments, I bet that you always think at least several moves in advance during a draft. If you use a draft queue to remind yourself of some players you're interested in in your league's "draft room," then you're definitely planning ahead.
Planning ahead is inescapable and helpful. Perhaps you "reach" for a particular second baseman because he's the last one left that you put any decent value on. That's planning ahead. Perhaps you decide to first draft a shortstop because there are still many acceptable closers left. That's planning ahead.
A form of planning ahead is called "backward iterating" or "backward induction." The above examples are fuzzy cases of backward iteration. Basically, you think about what you're going to do several steps ahead and that, in some way, helps you decide what to do in the first or present step. So much expert advice, if you listen closely, is a form of backward induction.
Here's the thing, though: You must work in the element of chance, of uncertainty. Whether or not you backward iterate, there's no getting rid of chance. But a common mistake is to put too much faith in your end-of-draft strategies, leading you to make early-round mistakes.
An example: You've done your valuations and compared them to some ADPs or consensus valuations. You see, perhaps, that you project Ryan Doumit to be a 15th-round value but that his ADP is (just making this up) in the 21st round. "Ah ha," you say, "I can wait on drafting a catcher and still get a good late-round value." Moreover, even though Victor Martinez is still on the board at the end of the sixth round (making him a good value, according to your projections), you wait on getting a catcher, prioritizing other positions first. You'll draft Doumit in perhaps the 17th round and still make an expected profit.
But what happens if by then someone else has taken Doumit? You may be left with Rod Barajas or some such character.
The problem is that your strategy was contingent on a very distant forecast—in this case, at least 10 rounds into the future. Ten rounds is likely an eternity in drafts—equivalent to forecasting the rain a month out or presidential elections 20 years hence. There's lots and lots of potential variance. Sure, you can probably forecast where a player will be drafted on average (ADP does a pretty good job of this, almost a fortiori). But strategies such as the one above aren't built on average; they are built on forecasting extremes. You really don't want to be too late drafting a position if you're backward iterating in such a risky manner.
The use and misuse of backward iteration casts a tightrope that we all must walk. On the one hand, drafting a "steals guy" in the second round just because you drafted a "power guy" in the first round seems excessively precautionary. Don't give up the chance to draft another "power guy" who may be the better value at that point; you'll be able to get steals later. However, you shouldn't put all your eggs in the Nyjer Morgan basket for the 10th round either, drafting only non-steals guys till then. Same goes especially for closers. Whether or not you subscribe to some version of the "don't pay for closers" mantra, planning on getting a specific closer in the late rounds is folly.
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.