When To Trust & When To Panic

 In Buy-low, Sell-High, Trading

There is nothing more frustrating than trading away or dropping an underperforming player only to see them explode after they are picked up by one of your opponents. It’s not a fun situation and it happens to the best of us. Sometimes this is due to bad luck. You need to get lucky at times when playing fantasy basketball and everyone goes through spells where the fantasy gods are not on their side. But sometimes it’s not due to bad luck. Sometimes it’s a preventable mistake. To make it easier to avoid such mistakes, I’ve put together an analysis that looks at when past performance becomes predictive of future performance. By knowing when player value starts to stabilize, we can avoid giving up on players too early and avoid buying high on early-season overachievers.

To do this, I looked at various groups of players’ rankings as of a week and then their ranking for the rest of the season. For example, I looked at the rankings of players as of Week 4 and then compared that ranking to their rest of the season ranking (Week 5 to the end of the season). For each set of players, I then calculated the absolute median of the differences between that group of players’ performances up to that point in the year and their rest of the season performances (i.e. If Player X is ranked 14th through Week 4 and then from Week 5 on they were the 29th most valuable player in fantasy, Player X’s difference is 15. I then did this for the top-25 players ranked by last season’s ADP and calculated the absolute median of the group’s differences). I chose to look at the median instead of the mean to avoid having the analysis spoiled by outliers.

This approach allows us to determine how predictive past performance is at different points in the season and how much a player’s value should be expected to fluctuate at different points in the season. As expected, how a player performs early in the season, especially during the first two weeks of the year, doesn’t tell us much about how a player will perform the rest of the year. Player performance becomes more predictive as the year goes on and by Week 12, you should have a pretty good idea of how each of your players is going to perform from Week 13 until the end of the season. It’s worth noting that past performance becomes less predictive of future performance late in the season. This is likely due to the shrinking number of future games and the late-season craziness that is the bane of fantasy basketball players.

Absolute median change by ADP:

I first grouped the players by ADP to see how player quality affects when we can trust players. Better players are more consistent on a game-to-game basis and the top players are extremely consistent. By Week 4, players drafted within the top-25 picks are usually performing within a round of their rest of the season ranking. Our analysis tells us that if a player drafted within the top-25 is ranked 12th after Week 4, then that player has a good shot at finishing within the top-25. However, if that player who is ranked 12th after Week 4 was drafted between picks 26-to-50, then you should be much more skeptical of their ability to continue to produce at a similar level. The predictive value of past performance decreases the later a player is drafted and you can’t really get comfortable with late-round picks or waiver wire pickups at any point in the season. This makes sense since late picks and waiver wire pickups are usually players with smaller roles that often change as the year goes on. They also tend to be forgettable NBA players and less talented players are more inconsistent than top-end players.

To summarize, for players drafted within the top-25, you will have a pretty good idea of how they will perform the rest of the year by Week 6. For players drafted around the third and fourth round, there is more variance, but usually, by Week 6 they will be playing at a level within two rounds of their rest of the season ranking. Players drafted in the seventh round or later will usually see their value fluctuate significantly all year long. This is especially true of late-round fliers.

Absolute median change by Position:

A player’s position doesn’t appear to have an impact on when we can trust a player. Power Forward performance was less volatile last season, but that was likely due to the large number of elite fantasy assets with power forward eligibility. Seven of the top-13 players in nine-category leagues had power forward eligibility on Yahoo last season.

ADP vs. USG%

I also took a look at how usage affects consistency. It turns out that high-usage players are very consistent, even if they are not particularly valuable fantasy assets. The top-5o players by ADP and the top-50 players ranked by USG% had about the same level of consistency throughout the year despite the top-50 players by ADP being much better on average. The average final per-game ranking of the first 50 players drafted on Yahoo last year was about 37. The average final ranking of the top-50 players ranked by USG% was about 55. This analysis shows us why the punt points strategy can be difficult to pull off. In that build, you will be targeting lower-usage players which will likely lead to some consistency issues for your squad.

Follow me on Twitter @AdamGStock for the latest fantasy basketball news and analysis. I will gladly answer any fantasy basketball questions that you may have. 

Recent Posts

Leave a Comment

Start typing and press Enter to search