Perpetual Leagues: Making Your Keeper List
by David Luciani
Published January 25, 2002

One of the most common emails we get from subscribers to our online ranking forms is in reference to the note at the top that reads: "You should finish the draft with a roster that costs about 75-80% of its value or you will finish in the middle of the pack!"  In a standard $260 Rotisserie league, a middle-of-the-pack team gets $260 worth of results.  A first place team pays between 75-80% of value and in fact, after analyzing hundreds of leagues worth of standard Rotisserie data, we have concluded that the optimum amount is actually somewhere around 78%.  That is to say that if you pay 78% of value, or in other words get $333 worth of players for $260, you will win a standard 12-team Rotisserie league.  The more teams that are in your league, of course, the more value you need to get.  Fortunately, the bigger leagues, with twenty-five teams and such, often allow you to choose from the same players as opposing teams, which means that ten guys in the league can own Pedro Martinez or Randy Johnson.  That opens up great possibilities for getting players at much less than their value but we'll deal with those leagues in a separate essay.

The focus of this article is to deal with the perpetual league.  The problem for most fantasy leaguers is exactly how to create a roster that only costs 78% (or 80% if you want to keep the calculations easier at the expense of a slightly weaker team) of its value.  One of the wonderful things about perpetual fantasy baseball is that finalizing your keeper list is the point at which you determine how much you'll be able to spend in the draft.  If you make a bad keeper list, you WILL lose your league but if you can retain players at much less than they are worth, you will be able to bid full price or even "overbid" for the players in the pool.

Let's create a real keeper list in a 12-team AL 4 category Rotisserie league and see what it enables us to do.  For subscribers to our forecast section, you can use our online ranking forms but if you're not a subscriber, use whatever set of values you're depending on to determine what you believe a player will be worth in 2002.  I'll use ours here for the example.

Let's say you're allowed to protect just seven players in your Rotisserie league and you've narrowed down your potential keeper list to the following players (with their 2002 salaries listed, if you choose to keep them):

Now, you know you want to find seven players from that list but you're not sure who to keep.  At first glance, the stars, such as Rivera and Ivan Rodriguez, might be the obvious choices.  So, let's go through step two.  Get the projected fantasy values from whatever source you're using and put those numbers beside the salary it would cost you to keep each player as follows:

There are really two approaches that you can take to build your keeper list.  One is what I will call the "Return on Investment" approach and the other is the "Maximizing Value" approach.  I prefer the latter but I've won leagues with both and so I will outline them here briefly.

The ROI approach is simply that you determine what percentage "interest" each player gets on your investment.  So, from the example above, Bartolo Colon gets you $23 of value for your $22 which is a 4.5% ROI.  Under the ROI approach, you protect the minimum number of players you can who get you the highest ROI.  If you protect too many, you might end up overfilling your roster with undervalued players but players who don't accumulate the stats that will win you the league.  Under the ROI approach, restrict yourself to the minimum number you must keep.

The second approach, and the one I have found preferable in recent years, is the Maximizing Positive Value approach.  This simply means that you subtract the amount a player will give you from what you are paying him and you protect the minimum number of players you are allowed unless adding any extra keepers is permitted and the extra keeper will be paid no more than 80% of his value.

So, using the Maximizing Positive Value approach with the above list, we create the following chart:

Under the Maximizing Positive Value approach, and if we were limited to protecting seven players, we would protect players in the following order: Escobar, Garcia, Lidle, Conine, Tyner, Cox, Abbott.  That's a total salary of $51 in salary and a projected value of $84.  In this case, if our league allowed us to protect more than seven players, we would add Robert Fick into the equation because he is being paid just 71% of his expected value.  In that scenario, we have an eight man protected list with $56 of salary and $91 of projected value.

Now, a few readers might ask the following: What if my list is so weak that I have players with negative expected value?  Well, it doesn't mean you can't protect them.  In fact, you might have a couple of players with such great positive value, such as Kelvim Escobar in the above example, that you can afford to keep them.  Having said that, you are always wise to maximize return and therefore, those are players that you should move for an undervalued player from another team.  Most other fantasy GMs will be aware of a bargain but you can often move a superstar, such as Mariano Rivera in the above example, for a $5 player who will be worth $10, even if the player's  not a superstar.