Dear Fantasy Speed Skating Managers,
It was a slow week at work for your stepdad, so I spent some intimate time with the Skater Predicted Points (Skater PP) Model and made some graphs, instead of working on frivolous things like improving road safety or inspecting crosswalks. Specifically, I’ve looked at how my and dad’s predictions are doing so far. As expected, some are going well and others are in worse shape than Christian’s team management.
Skater Rankings
First, I compared the Skater PP Model’s output for each skater to their performance so far this season. In theory, this shows which skaters are “overperforming” and “underperforming” compared to the model.
I would have loved to also test dad’s predictions, but I don’t have his skater-level data, if such data even exists. I have no clue how that hack comes up with his rankings. In fact, I’m pretty sure that he ranks teams based on vibes and then sends me a spreadsheet with a bunch of random numbers in it to keep me confused.
Anyways, Kimi Goetz and Mei Han are the two skaters who most exceeded the model’s expectations. In the league’s tradition of absurd beginner’s luck (I’m looking at you, Ryan and Greg), Peter has both these skaters on his team. Masaya Yamada, Qishi Li, and Tatsuya Shinhama are also notable overperformers. As your real dad discussed, several managers picked up on Yamada’s strong performance at World Cup 1 and unbenched him for World Cup 2, where he earned exactly zero points. If this is a slap to the face to managers who unbenched Yamada, it’s a rusty baseball bat to Christian’s delicate shins (like seriously, can you imagine dropping Mei Han to pick up Yamada!?). Greg was unfortunately the only manager to skate Yamada in World Cup 1 and benefit from this overperforming late-round pick. Like everyone else, stepdad is hoping that Greg runs out of luck soon and returns to the league’s Emin-chasing basement where he belongs.
Some skaters have done almost exactly as well as the model predicted. Gao Tingyu was the closest, followed by Ayano Sato, and Gabriel Odor. All three of these skaters have a “Z-Score Difference” of less than 0.01 between the model and their actual number of points. Congratulations to Eric and Emil for bringing Gao off the bench for World Cup 2 and being the only team to harvest his points. (Side note: I might be committing statistical atrocities by comparing Z-Scores between the model and actual points… if someone wants to school stepdad on how to properly do statistics, please go ahead).
Finally, some skaters are drastically underperforming the model’s predictions. The worst offenders are Min-Sun Kim, Vanessa Herzog, Jordan Stolz, Laurent Debreuil, and Michelle de Jong. The first few picks are critical to a team’s success, so managers with any of these skaters on their teams should be praying for a turnaround. Cody and Yankun should get down on their knees and pray extra hard, as they’re suffering from having two of these underperformers on their teams. Managers were clearly smarter than the model regarding de Jong’s performance as she was only picked once (by Darby) in the 9th round.
For those wondering how their picks’ performance compares to the model, here are some graphs:
Battle of the Dads
Many of you are probably dying to know how the accuracy of dad’s and stepdad’s power rankings compare. With a single can of Lucky Lager on the line, the stakes are sky-high. Stepdad’s rankings are based on a hairbrained, reductive, and downright sloppy model. Dad’s predictions are probably based on a combination of astrology, exclusive access to World Cup skaters, and an anti-oil political bias. Also, dad cheated because he released his power rankings after World Cup 1 (at least he admitted it in his previous column).
The good news is that both dad’s and stepdad's rankings are positively correlated with the current results, meaning that neither of your dad’s completely dropped the ball (unlike Christian- did you hear that guy ditched Mei Han to pick up Yamada!?). The bad news is that dad’s power rankings are currently doing much better than stepdad’s. We’ll see how this plays out over the season, but it looks like stepdad might be buying one (1) refreshing can of cost-effective beer for your real dad.
For those who care about the numbers, stepdad’s power rankings have a Spearman’s rank correlation coefficient of 0.35 with the current team rankings. Dad’s rankings have an impressive Spearman’s rank correlation coefficient of 0.67. For those who care deeply about the numbers, Dad also outperforms stepdad if using Kendall’s Tau or Sommers’ D as the test statistic. All results are significant at the <0.02 level.
The following graph shows which teams are underperforming and overperforming according to both power rankings. Essentially, the graph shows how proud or disappointed dad and stepdad are of each manager. Sofia is currently the most disappointing manager, mostly because Stolz didn’t skate in World Cup 2. Thomas and Cody are also doing their best to disappoint stepdad. Dad is generally less disappointed than stepdad, but he’s still dismayed by Jake and Dylan (who also got clobbered by Stolz’s absence from World Cup 2).
Sitting in the middle are Cooper and Molly/Clair/Victoria, who are performing almost exactly as both dad and step dad predicted. Good work gang. Someone has to be perfectly average.
On the flipside, Daniel is crushing stepdad’s expectations (but doing about as well as real dad predicted). Jill is doing much better than dad predicted, which is both a result of Shinhama skating good and dad having almost no faith in her at the start of the season. Regrettably, Greg and Ryan are both outperforming both dads’ expectations. Your dads can't agree on much, but we can agree that we don’t want to see Greg win the draft, or see Ryan win again.
That’s all for now, everyone. If you have an issue with the rankings or want to rake your stepdad over the coals in the comments section, please go for it. If you haven’t already, please fill out the Poll of the Week, which will help me with my next column.
Cheers,
Stepdad