With the 2020 season starting we crunched the numbers and put together the Morning Chalk Up’s Power Rankings. These rankings are intended to both reflect an athlete’s previous performances as well as forecast success in 2020. Because these rankings are meant to forecast, not all past performances are treated equally and extensive modeling takes place. It is important to note that these rankings are purely numbers derived, and we acknowledge this method will have its shortcomings. For example, Jacob Heppner is ranked 27th for the men, and while we may personally believe he will finish better than 27th at the Games, there is still limited data for us to model from to give him a higher ranking (compared to athletes who had 2017 and 2018 seasons).
For these power rankings we considered:
- 2019, 2018 & 2017 CrossFit Games
- 2018 & 2017 CrossFit Regionals Worldwide leaderboard (all results merged)
- 2019, 2018 & 2017 CrossFit Opens
- 2019 Sanctionals (15 total tournaments)
To learn how we should forecast Games performance, we reviewed how the 2017 and 2018 seasons predicted the 2018 CrossFit Games. Ultimately the power rankings looked like this when forecasting the 2018 Games.
Power Ranking Points vs 2018 Games Place
* If an athlete was injured and withdrew, think Sara Sigmundsdottir, then they were not considered in our modeling.
We took the lessons learned and did our best to replicate and expand them to fit the new 2019 Sanctional season. The major ideas were:
1. Unsurprisingly, consistent elite performance is the best single predictor of future success. However, simply averaging past performance does not do a great job of isolating underlying ability and predicting Games performance. This is because an athlete’s weaker performances can be discarded and only their top performances considered. Especially if the athlete has an extensive competition history. Many things like injury or illness, choosing not to peak for a certain event, an unfortunate workout in a condensed competition, or similar situation can derail an athlete. Going into any competition this model will assume that all athletes are peaking and bringing their best. Therefore, only an athlete’s top 4 performances in the events previously listed are considered.
2. Live competitions (think Regionals or Sanctionals) have historically been a much better predictor of Games performance compared to the Open.
Similarly, prior Games performances were a better predictor than recent Open performance. In these power rankings it is assumed that Sanctionals are more similar to Regionals than the Open, as they are in-person competition and all that implies.
3. Not all Sanctionals were equal. Some competitions, like the Rogue Invitational and Dubai CrossFit Championship, were loaded with talent. In these a 5th place finish might be as impressive or more impressive than a win at a Sanctional with a lesser overall talent level.
The talent level present at a given Sanctional was considered and used to scale the points available.
4. The separation between finishes is not always the same, meaning the performance variance between 3rd and 10th may be different at different competitions. For example, the women’s division at Fittest in Cape Town only had 15 competitors with maybe 8 capable of competing at an elite level. Conversely, nearly all if not all the athletes at the 2018 Games were capable of elite performances. Further, in the merged Regionals athletes with a rank 100th and beyond were worthy of consideration for points. The point spreads between placements and the depth of points awarded varied accordingly.
5. The final variable considered was how recent the competition was. This caused the points available from the 2017 and 2018 Seasons to be slightly reduced when predicting the 2020 season.