An important part of drafting is knowing which positions should be weighed more heavily because of positional scarcity. That is the only reason why catchers like Victor Martinez or Brian McCann get selected in the 3rd round with the sluggers. They most likely will put up worse numbers overall than their non-catching counterparts (i.e. other hitters taken in the third round) but still provide equal value because of the weaker catchers your opponents will be forced to choose from in the later rounds, while you can get better production from those leftover first basemen.

Obviously the crux of this whole operation is being able to determine which positions should be deemed scarce and which should be deep. First of all, this has to be a projection-based operation. Using the stats of last year won't do any good because the depth of positions change every year. People return from injury, people retire, and people break out; if you cannot at least attempt to take into account future production you will always be living a year behind. A simple spreadsheet of Marcel projections--or any other projection system's projections--will suffice.

So now that we have the projections all ready to go, but what do we do with them? A logical method seems to be to take the average wOBA (or OPS or WPA/LI, whatever) of all the players' projections in a given position and rank them in order, form best to worst. There you go, there are your scarcity rankings.


Hold on, there. You are forgetting something important. Average does not tell you the entire story. Consider this scenario:

There are two groups of three hitters. In group A the three batters' batting averages are 0.280, 0.300, and 0.320; and in group B the batting averages are 0.200, 0.300, and 0.400.

Both of the averages (means) of the (batting) averages are 0.300, but if you were to imagine the two groups of hitters are two different positions to draft from, it becomes apparent that selecting a player from group B is more valuable than one from group A.

Thus, it is also important to look at how the projections are distributed around the mean. It is a normal distribution or uneven distribution.

If two positions have relatively similar average levels of production, then taking a look at the distribution can be an important deciding factor. If the distribution involves a few outlying players far above the mean, a sharp drop off, and then a gathering of players slightly below the average, then drafting one of those few outliers should be a priority. And vice versa, if there is a significant bunching of players just above the average, and a subsequent grouping of players well below the average, then perhaps waiting to take a player from that position is ideal. The first player taken might not be much better than the tenth. Make sure to recognize where that sharp drop off is though, and make sure you err on the side of caution so that you do not end up having to select after it.

This is a similar concept to tiered rankings where players are not just ranked in order but also in separated grouping of players of comparable ability. Generally speaking, it is desirable to select players toward the bottom of their tiers because they put up similar numbers to players taken ahead of them. Of course the risk exists that all of the players in a tier get selected and you are forced to select from the next lower tier.

Overall, distribution of talent throughout a position is something important to take into consideration when it comes to the issue of position scarcity. Catchers may be relatively bad, but if they are all similarly bad then it makes no sense to draft one of the other. I'm not saying that this is the case with catchers this year for I have yet to run the numbers for 2009, something you can expect in future column.