The Process by which I outline here for evaluating players, while accurate, is not to be used for the draft day. Doing so would result in an enormous misevaluation and could be detrimental to your season for several reasons. First of which, is Positional Replacement Evaluation
Positional Replacement Evaluation works on the idea that all players in the draft pool are valued against the expected least valuable player drafted at their respective position when the draft is over. Traditionally, this is accomplished by multiplying the number of teams in the league with the number of players needed per team per position.
In a 12 team league where 1 catcher is needed per team, after the best 11 catchers have been drafted, there is no real advantage in the traditional model to take the next best catcher available as all the competition has already got their catcher situated. The z-score value of this player is then referred to as the Catcher Replacement Player. In theory, you would be able to go the entire draft, not select a catcher, and this player would be at the top in the free agency pool.
By this method, this player has zero value among the starters in the league as every other team's starter is expected to be better. The traditional model in turn, takes the z-score value of this replacement player and subtracts it from every catcher in the draft to make a Positionally Adjusted z-score.
Usually, this results in taking value away from players, but sometimes it can add to a players value. In reference to the example, if this 12 team league had used 2 catchers per team instead of 1, that would mean the 24th catcher in the league would be the replacement player. Typically, catchers going this far down in the league will have a negative z-score because the offensive talent is typically lower than other fielding positions. So when it comes time to adjust for Positional Value, you would be subtracting a negative number, which is the same as adding even more value to the Joe Mauers and Buster Poseys of the drafting set.
Many people will refer to this as positional scarcity, and I feel is an effective way of achieving a model evaluation.
However, I feel there are some oversights in this method. The are a few problems this method faces. Two of which, that I will explain later in this article, is the issue of dealing with multiple-position eligible players and UTIL positions. The first however, is dealing with bench spots.
Identifying the Replacement Player
When putting together my very first algorithm I quickly realized that bench spots were going to be a problem. I had a hard time deciding if they should be taken in account, and how to go about doing that. Bench players have value in a fantasy baseball draft - they are going to be drafted after all, but should it just be looked at taking the best players available for bench spots? When doing simply best player available you will have an imbalance in positional values that would not be representative of what a draft would look like. The league isn't going to take all 1B and 3B players as their reserve spots because they are the most valuable, no. There is going to be a balance of positions to achieve proper backups.
So should we just take an arbitrary number of players over the current replacement player, like 1.5x the amount of catchers needed in the starting spots. This could be close, but it is highly suspect that this rule would be followed exactly as there usually arent that many draftable positions, especially in leagues with a very shallow bench number.
Finding the replacement number within a position is critical to positional adjustment method and if we take a player much lower than what would ever be needed, like taking the 18th catcher available if we were to use the 1.5x rule in a league with 4 bench spots, catchers would be artificially inflated in value as it is highly unlikely 6 reserve spots would be utilized for backup mediocre catchers.
The Fantasy Baseball Calculator Spreadsheet uses a simple and more logical means to determine positional replacement values. What it does, is it will look at the number of games required to fulfill the conditions to meet the maximum number of expected to start per position, which is usually set to 162 in most formats. The calculator will find the best player for each position that is not needed to meet the games threshold. In a 12 team 1 catcher league, the number of games required would be 1944. As it is the first 1944 games from the top players in the league that can be scored, those have intrinsic value within this setting.
Looking at 2012 stats, it took 16 catchers to exceed this number. Even though the draft and the settings of the league may not accomodate this space, there will be 16 catchers that should provide value to the end of the year stats. The method of positional adjustment then follows the traditional means of subtracting this value from the rest of the pool.
Dealing With Multiple Position Eligibility
The second foremost problem I have found with finding replacement level players is when dealing with utility players and multiple position eligible players. To deal with these issues, the Fantasy Baseball Calculator will look at the Replacement Values for each position from above. For players with multiple eligibility, it will identify which one will take the least value away from the original z-score and use this as the adjusted value. This player is then viewed by the program as only this new position, thus allowing a more accurate replacement level for the position it was taken away from to be used.
Let me use a small example to explain this - I realize when typing it out how poorly I am conveying the message.
Example: it has determined that the 12th shortstop and the 12th 2nd baseman are to be used as their respective replacement values. the sixth ranked shortstop and the 4th ranked 2nd baseman is actually the same player due to his multiple eligibility. In effect, this means that there are really only 23 players in a pool that is viewed upon as being 24 players deep.
in this setup, the 2nd baseman position has a lower replacement score, so this multi-eligible player is best utilized at 2B. This means we have 12 second baseman and 11 shortstops. After sorting this player out, the Fantasy Baseball Calculator will then recount the number of games it takes to meet the threshold again. At the very least it will be 1, but it may even be 2 or more depending on how few of games the next best players have played. After the threshold has been met, the replacement value is assigned and then the normal model follows.
Generic Position Eligibility
Generic Position Eligibility can be a difficult problem to tackle. In a league with a UTIL position, you have a position that anybody can fill. There are two main problems I have seen with trying to properly evaluate this position. For simplicity, assume a 12th replacement level position for UTIL spot.
The first incorrect method is to take all the players that did not exceed the replacement value for their position and take the 12th best as the replacement level position. If you subtract this replacement level from these players, you will have situations where the 13th best 1B will be valued more than the 12th best 1B. The 12th best is already adjusted to zero. If the 13th best 1B has a value greater than the 12th replacement UTIL position, when you subtract that value, you will have gained a positive number and thus, incorrect value assigned.
A common way to fix this problem is to add the amount gained to each position by this method to the already adjusted values in the standard positions. However the problem with this is it inadvertantly adds values of production to the player pool which, especially when valuating for Auction formats, will reduce the ratio by which the top hitters have an advantage thus lowering their value, and similarlity raise the value for the lower hitters in the ranks. This method distributes too much even money over the pools and thus should not be used.
The way the Fantasy Baseball Calculator spreadsheet solves this problem is by identifying all the players that have not surpassed the replacement level for each position. The number of games it takes for the UTIL max games to be satisfied is then identified and the player that meets that condition is identified. The Fantasy Baseball Calculator then will take this replacement score, and compare it against the singular position replacement score. So in most formats, the C replacement score will be lower than the UTIL replacement score. For all catchers, the lower C replacement score is used. For 1B, however, in most formats the UTIL score will be lower than the singular 1B replacement score. In this event, the UTIL score will be used. The calculator functions like this for all positions, first looking at singular position, OF, MI, CI, INF, and finally UTIL. Whatever is the lowest replacement score for any player eligible at any of these generic positions, the calculator will adjust with that value.
Now with all of that said, Click Here to let me explain to you why using this ranking is, once again, completely wrong.