I got one of ‘those’ this year because I was bored at the train station, total garbage.  They have a mock draft that’s from November and terrible projections.  They don’t tell you any formula or system they use for them, you get the impression a couple guys go yea Delgado is gonna hit 40 this year, write that down!!

If we accept that “everyone has access to a number of free projection systems, all of which have roughly the same level of predictive accuracy” then doesn’t it follow that everyone has access to a number of free draft strategies?  I mean, surely if you’re the kind of person who knows of and seeks out Marcel/CHONE/ZiPS/whathaveyou, you’d be reading the Dierkes’ and Carty’s of the world too, no?  And at least from what comes across my monitor, it seems like year after year, these strategies are becoming more and more similar, presumably as inefficient ones are weeded out.  Are there actually reputable fantasy analysts promoting taking closers early or ignoring peripherals?

And then, if we accept this (which by all means you are free to not do), does the question not become a much simpler chicken/egg question?  Optimal strategy inherently relies on proper evaluation, or else things like LIMA and 10-20 Steal Guys aren’t helpful at all.  On the flip side, with optimal evaluation, any strategy is available to use, and indeed will make those strategies even more successful.  This would be similar to Mr. Podhorzer’s answer, and seems somewhat undeniable to me.

I know it’s a bit, shall we say, semantic, to boil it down this way, and as such takes the fun and debate out of the question.  But if we’re really answering it, I’m not sure it’s disputable: without the evaluation, the strategy becomes worthless.  Now, as I said at the beginning (or didn’t say, I suppose), if you think everyone’s evaluation methods are reasonably similar yet strategies differ significantly, then ok; strategy becomes the trump card.  I just don’t believe that to be the case.

Dan -
Agreed that Marcel/CHONE/ZiPS projections are only known to hardcore baseball / fantasy baseball fans.  While sites like ours, Tim’s, this one etc. will reference these sources (or create premium projections like THT), the bigger sites like ESPN or Y! won’t expose the average online player to these projections.

In my analysis this year of Marcel/CHONE/ZiPS (and some feedback from their creators), I found a exploitable flaw.  The player AB/IP for these systems is generally overestimated as it is not tied to projected playing time.  You can get an edge by converting this data into rates (e.g., HR/AB) and using updated AB/IP either through your own projections (leveraging public depth chart info) or those of a site like Baseball Prospectus (fantasy subscription - PECOTA has same issues as Marcel/CHONE/ZiPS) and re-projecting the totals.  Changing someone like Wieters from 600 Plate Appearnces (PECOTA) to 400 (latest BP) will dramatically reduce the value of the counting stats.

Tim -
Totally ruined those magazines.

Last point…no doubt that drafting 5 category guys is desirable but there aren’t many out there.  There’s a definite bias towards HR/SB guys - especially those that are weak in R, RBI, and/or AVG.  In leagues like this, I can see 3-category players being undervalued.  There’s no reason every player on your team needs to be balanced as long as your team is.

The answer to the question really depends what kind of league you play in.  In weekly transactions leagues I think player evaluation is a reasonable answer.  Any leagues with daily transactions, strategy is far more important.  And in formats where you can turn over your entire roster every day, the ability to take projections and adjust for the context of that day’s game is 99% of winning.