Inflection Points: Steady As She Goes
Previously, I introduced a statistic called inflection points which measures how frequently a player either greatly exceeds or falls short of his average points per game in consecutive weeks. More than just a measure of consistency, inflection points attempt to tell you how often a player puts you in a situation where you are likely to make a mistake with your roster, either by benching a player you should start, starting a player you should bench, or trading a player below his real market value.
In my mind, inflection points are meaningful in and of themselves. Of course, you cannot really forecast when a player will get hot or cold—or, rather, I won’t try to—but I believe that if you want to decide between two players with similar profiles whose major difference is their number of inflection points in recent seasons, you should take the one with fewer. Even if two players will produce the same number of fantasy points, the one whose value you can more accurately assess is the more valuable one.
That said, in my previous post, I never did disprove what really was the impetus of the article in the first place, which is my contention that it is a mistake to dramatically re-evaluate your opinion of a player when an inflection point occurs. I will try to do so now.
If a player has two bad weeks in a row, you are tempted to drop him in your rankings, and if a player has two good weeks in a row, you are tempted to raise him. I decided to look at a player’s production before and after inflection points to see if there was any reason to do so.
Since I already knew which players had consecutive bad or good weeks and knew when those occurred, I simply calculated fantasy points per game before and after those consecutive weeks so I could compare them. I included only players who had inflection points between weeks 6 and 12 of the last two seasons so that they had at least a four-game sample on either side to calculate points per game. It still isn’t perfect, but with a limited sample to start with, it seemed like the best compromise.
First, here are points per game before and after negative inflection points:
And here are points per game after positive inflection points:
Not only are our instincts incorrect, they are actually counterproductive. Over the last two seasons, players at every position have scored more points per game after a two-week slump than before it and fewer points per game after a two-week spike than before it.
Clearly, there is something of a selection bias in that players that fell apart after a negative inflection point would not have qualified as a star in prime, star in decline, or RORA, which I used as a filter for inclusion. Of course, that is also very much the point.
Players with established track records of production are established. Most of them will have occasional peaks and valleys. You should not allow them to change your opinions on the players. In fact, given that most owners are sensitive to those streaks, go ahead and buy established players that slump and sell established players that streak. If recent history is an indication, you’ll come out ahead.
Find more of our Dynasty Content here.