Article:Fantasy Insight: Category Scarcity

Opinion Top}} With fantasy baseball draft season among us, I've been laboring endlessly over cheat sheets, projections, ADP lists, and "expert" analysis in an attempt to put together that all important customized cheat sheet for my own draft day.

Although I do not publish alot of material regarding stats and sabermetrics here or at my blog, I am a stat lover. I love Baseball Reference and Baseball Prospectus and all of those other obscure statistics outlets. I download the Lahman Database every year, and I've got a queue line of about five books which focus almost solely on the statistics of baseball that I'm anxious to read, including Bill James' Historical Baseball Abstract.

It's really quite sad, but the reason I do all of this number crunching is, in all honesty, to gain a competitive advantage in my fantasy leagues. I try to find or develop a stat that forecasts perfectly and tells me which guys will outperform others with 100% accuracy. Don't get me wrong. I do enjoy the statistics of the game independent of fantasy baseball, but I will never be mistaken as the next Bill James. So, I figure that I may as well point my efforts towards something productive and useful to me personally. And what better way to use statistical analysis than for fantasy baseball, right? For those of you who play fantasy baseball, you know that there's no better feeling than claiming the league bragging rights or whatever other monetary prize may be on the line.

With that in mind, I'd like to present a potentially advantageous fantasy strategy that I've randomly discovered while putting together my cheat sheets and projections. (Okay, maybe I didn't "discover" this idea in the true sense of the word, as I'm sure someone out there has beat me to it. If so, consider this post supporting evidence.) You've heard about the whole position scarity theory, I'm sure. But how about the theory of category scarcity?

Many fantasy sources seem to stress stolen bases and saves as areas of emphasis for a fantasy roster. I've never really understood the premise behind these assumptions. But, in my analysis, I believe that I have confirmed these suggestions.

Here's what I did. I compiled Major League totals from the past three years for each of the standard 5x5 categories. (I ignored rate stats - i.e. batting average, ERA, and WHIP - since those are a little more difficult to manage, calculate, and project.) Then, I compiled three-year totals for each player in these same categories and compared them to the Major League totals to get a percentage. For example, there have been 94,488 strikeouts over the past three seasons. Josh Beckett has struck out 518 of those hitters in that time span, a percentage of 0.548%. My goal with this whole process was to somehow find out who the best five-category contributors have been over the past three seasons.

What I found out, though, was this idea of category scarcity. The three-year leaders for both stolen bases and saves had far higher percentages of the league total for their respective category than any of the other category leaders did. The table to the right explains this further; it displays each of the seven non-rate stats standard for a 5x5 league, the accompanying leaders, three-year league totals, and percentage accumulated based on those league totals.

So, how does this actually explain category scarcity, and how does it apply to a draft strategy? Well, without going into all of the specific details and processes, I basically calculated the sums of each of these categories from the past three seasons as they pertained to my particular fantasy league. In other words, I calculated the three-year outputs from my fantasy league for stolen bases, home runs, runs, RBIs, strikouts, wins, and saves. (And, yes, I did have these figures readily available.) Then, I recalculated the percentages accumulated for each player based on the three-year totals from my fantasy universe rather than the three-year Major League totals. Using the Beckett example, there have been 30,888 strikeouts in my fantasy league over the past three years, and Beckett has struck out 518 guys over that span, a percentage of 1.68%. By making this change in calculation, nothing really changes in a relative sense. The percentages simply go up. But by basing the percentages on fantasy totals, the results can draw a clearer picture of how scare the scoring categories are.

Ponder this. Based on the three-year totals from my fantasy league, the top 10 stolen base leaders from the past three seasons account for 34.29% of the stolen base total. Even more dramatic is that nearly 44% of all saves have come from the top 10 closers. In no other category does the top 10 even account for 25% of the total. Just for reference, it would take the top 21 home run hitters to reach the 34% mark and the top 28 to reach the 44% mark.

Realizing that being only average won't win any fantasy baseball leagues, I created a target threshold for each category that was simply the average of the previous three years' combined totals. With this threshold in mind, I could then determine an expected contribution from each player in each category based on their historical percentages relative to the fantasy universe. For example, Beckett's 1.68% strikeout ratio would translate to roughly 172 strikeouts if my league's projected strikeout total is 10,296. If my target threshold is 858 strikeouts (10,296/12 teams), Beckett's production would get me 20% of the necessary output to be average.

Let's revert back to stolen bases and saves for a minute. Based on my thresholds,  stolen base leader, Jose Reyes (5.1%), could account for 61% of the stolen bases necessary to achieve that particular threshold. Similarly, top saves man, Francisco Rodriguez (5.1%), could also account for 61% of the necessary save total to achieve that particular threshold. For comparison's sake, the top home run (Alex Rodriguez), runs scored (Jimmy Rollins), RBI (Rodriguez), strikeout (Johan Santana), and wins (Beckett) guys over the past three years would only account for 25%, 16%, 18%, 28%, and 26%, respectively. So, at least in theory, it appears that the fantasy experts are on to something when they suggest to focus a larger fraction of your efforts on acquiring stolen bases and saves come draft day.

The lesson learned in all of this mess? Stolen bases and saves are far scarcer categories than any other when it comes to a standard 5x5 fantasy baseball league. By being cognizant of this trend and drafting wisely, you should be able to remain highly competitive in these areas with minimal sacrifice elsewhere. You can reach your target thresholds for stolen bases and saves with as little as two guys where it would take as little as five to reach your threshold in home runs, for example.

I'm not suggesting you neglect the other categories altogether. In fact, neglecting home runs can be extremely damaging since with each home run comes at least an equal number of runs and RBIs. What I am saying is that it may not be such a terrible idea to draft someone like Jose Reyes with the top pick rather than ARod. Stolen bases and saves get sucked up extrememly fast by only a handful of guys. Meanwhile, if you miss out on the ARods or Prince Fielders or Ryan Howards of the world, there will still be plenty of home runs and RBIs to go around.