Making Projections into Values: Modifying Projections
Before talking about turning projections into player values in earnest, I should touch on the subject of modifications you should make to your
projections to make the values you derive more useful. While not strictly necessary, this can improve the effectiveness of your derived values.
The primary such useful modification takes into account games played. Keep in mind that this is not a necessary modification however, and it can
be somewhat complicated, so if after glancing at this page you don't feel up to it, feel free to move on to the next chapter in the guide.
Often, you will project that a player will play fewer than the maximum number of games in a season. A player might be currently injured, but is
expected to return during the season, or a player might be carrying a lingering injury which you expect to resurface, or a player might have a
history of getting injured or not playing through minor injuries. Regardless, a player who is projected to play fewer games should rightly be
projected to collect fewer positive statistics.
That doesn’t tell the whole story however. Often, if a player misses time, you will be able to find out about it, and play some replacement instead.
The replacement player might not be as good, but you will be able to get some kind of value out of them while your principal player is on the bench or the
disabled list. The bottom line is that you would much rather have a player, for example, that gets 5 touchdowns and 800 receiving yards and in 8
games played than a player who scores 5 touchdowns and 800 receiving yards in 16 games played. A way to quantify this difference is to modify
player projections taking games played into account.
The idea is to project the per-game statistics for an imaginary replacement player, multiply these by the number of games you would expect to use a
replacement player, and add these to the statistics you have projected for the primary player. In this way, you get a better sense for what you will
actually get from a player who should miss time.
Your ability to get replacement games for an injured player is affected in large part by the fantasy sport you are playing and by the settings for your
league. First, you should recognize that if a player tends to miss games, you may not always know beforehand when a player will not play. Often a
player is listed as day-to-day, and until the beginning of the game nobody knows whether this player will be active for the game or not. In a sport
like basketball or hockey, where it is relatively easy to make up missed games (because teams play every third day or so, allowing you to rotate
bench players into your starting lineup whenever their team plays), this may not be much of an issue, but in a sport like baseball or football you
may not be able to make up the games by the end of the season. Also, if a league uses daily transactions, you can quickly respond and bench a player
when they get injured and will miss time. In a weekly transactions league, you may miss many more games because of the delay in reaction time.
Take these factors into account when you are calculating how many replacement games you will be able to use if you draft a player who misses
Based on league settings and the sport in question, I usually guess that I will be able to use a replacement for between 70% and 100% of the games a
player misses. Of course, this presupposes that you keep reasonably close track of your team, and of injuries that occur. If you don’t plan on being
very active in following fantasy sports news, this percentage could drop to 50% or 30%, or even 0% if you plan to never check on your team at all.
Once you have calculated how many replacement games you should receive, you have to determine what kind of per-game performance you should expect out
of a replacement player. This is a difficult question to answer before you have completed your modified projections, but it is necessary. You’re
not necessarily looking for the performance of a waiver-wire level player, as you will probably be using one of your bench players as a replacement
player. You also want to make sure that you correctly account for position – the quality of replacement player at one position is likely going to be
different from the quality of replacement player at another position. I would probably do it this way – take a set of player rankings that you trust
(and that ranks a large number of players), and figure out how many players will start for one of the fantasy teams in your league at each position.
For example, if there are 12 teams and 3 starting guards per team, then there will be 36 starting guards – maybe you figure 20 point guards and 16
shooting guards. Take the 10 or so players in your ranking who are just below the starters – in our example, point guards 21 through 30 and shooting
guards 17 through 26. Using your projections, find the per-game statistical performance of each of the shooting guards, and average them to get your
replacement shooting guards stats. Average the per-game performance of the point guards to get your replacement point guard stats. The resulting
average is a decent approximation of what a replacement player should produce for your team at those positions. Use a similar technique to calculate
the production from a replacement player at another position.
Once you have estimated how many replacement games you should be able to use, and you have an estimate for the per-game performance of a replacement
player for each position, it is fairly straightforward to modify the total projection for each player. In each category, simply add the stats that you
project you will collect from replacement players for the games that your primary player will miss.
It is important to recognize that there is a hidden cost here – namely that by drafting and using a player who will miss time, you need to reserve a
bench spot or a disabled list spot for this player. Bench spots have a fair amount of intangible value, as you can and should use them for in-season
add/drop activity. Having a player or two who you have to hang on to but who do not provide current stats can hurt your flexibility to pick up
promising waiver-wire players. It is difficult to quantify this cost, but it does exist and should be remembered when it comes time to draft a player
who has substantial injury issues.
Next Section - Making Projections into Values: Fantasy Point Scoring