Featurette Friday: Predicting the Playoffs
Discover the Best Reader Questions, Article Flashbacks, Hockey Analytics Basics, Powerlifting Principles, and Financial Wisdom
Greetings, everyone, and welcome to this week’s Featurette Friday post.
Every Friday, we'll bring you a collection of small write ups.
We'll start with reader questions, which you can submit via Twitter @TheFiveHohl, my personal Bluesky, and our Facebook page. Follow us both on Bluesky and Facebook.
Next, we'll review and analyze some of my past work and opinions.
READER QUESTIONS
Twitter Questions
https://twitter.com/AavcoCup/status/1775189851366908407
Neal Pionk signed his four-year, $5,875,000 AAV contract in August of 2021 after filing for arbitration. After two terrible seasons in New York and being part of the Jacob Trouba trade, Pionk put up two strong back-to-back seasons in Winnipeg.
Here’s the thing. In 2020-21, Pionk played a fair bit with Dmitry Kulikov, Josh Morrissey, and Luca Sbisa. He also played a few games with each of Ville Heinola and Sami Niku that season.
The next season, his contract year, Pionk played heavy defensive zone deployment with the toughest line matches and came ahead. He also played every game but two with Derek Forbort as his only defensive partner.
The two games away, Forbort played with Tucker Poolman — and they stunk — and Morrissey played with Pionk — and did okay.
No one will convince anyone that Forbort is some elite blue liner, but he was capable, and his (and Kulikov’s) chemistry with Pionk may have elevated his play more than his actual true worth.
https://twitter.com/WPG_Chief/status/1778459444868735455
Yes, but no.
Goal differential is a lot better at predicting postseason success than the regular season; however, this is mostly because the team with superior goal differential is the team most likely to have home-ice advantage.
On its own merit, after accounting for home-ice advantage, 5v5 Corsi is still the strongest individual metric in predicting future outcomes, even in predicting the playoffs. BUT, no one is really limited to a single metric.
This tweet was in reply to one of my premium-tiered Wednesday posts, where I showed that Corsi is still a stronger predictive measure than expected goals despite expected goal differential models improving(and data becoming more accurate).
However, when I repeated the same test with weighted shots (the same model you see above in Pionk’s defensive partners) which uses goals, expected goals, and blocked shots (i.e., the Corsi shots not already in expected goal models), it is far more predictive than all of them.
You also build a better playoff prediction model if you use both 5v5 and special teams metrics, but keeping them stratified.
https://twitter.com/TheS1lverScrub/status/1778582409434513816
There are always logical reasons. Whether or not the reasons are correct is a whole other issue. I don’t want to rewrite our last Monday’s post, but we know that Kyle Connor and Mark Scheifele, both this season and the previous six seasons, tend to do poorly in any shot metric, including goal differential. That is, except when with Nikolaj Ehlers.
There are reasons why Bowness likely prefers to do what he’s been doing.
He definitely seems to like having right-hand shots on the right wing as much as possible. There’s some merit to this as wingers playing on their off-wing do have a higher shooting percentage but a lower shot differential and save percentage (holding all other things equal). So, it’s suboptimal… but only if all other things are held equal. The talent and performance discrepancy between Ehlers and the alternative wingers on the top line is quite substantial.
He prefers to spread his players into roles, so he prefers each winger to play a bit of a different style. Again, there is some research-backed merit to players covering for another’s weaknesses. But, there’s also research-backed merit showing that doubling/tripling up on skillset/archetypes can also be even more advantageous. For example, two puck moving defenders on average will still outperform one puck mover and one defensive defender (holding all other things equal).
Coaches don’t like unforced errors. There is again some merit to this. Part of their job as coaches is to try and coach out unforced errors because that’s an area you can improve. Also, coaches don’t want to look like they are playing favorites to those who make more obvious and anti-system mistakes. I’ve always said that a coach will rather play a 6/10 player playing at their full potential than a 9/10 talent playing a 7/10 game. I don’t fault Ehlers’ effort, but I can fault his gaffes.
There’s more, but there’s a lot to heuristics and biases that make them somewhat understandable and logical in a vacuum.
It’s all connected to losing the forest for the trees. And it makes sense because the coaches are in the forest and they don’t get a bird's-eye view like outsiders do. It makes it very, very difficult to remove the preconceived biases and such. Especially when focusing on the trees is at times their primary role.
The truth is that Bowness does not run a meritocracy, but it’s an understandable mistake.
https://twitter.com/WPG_Chief/status/1778582156949782997
I have to ask, does Ehlers actually struggle on the top line? Like he did in the recent game, but in general the top line scores more, gets scored on less, creates more chances, and allows fewer chances with him on the line. Is he actually struggling?
Now, he does make more weird decisions on the top line, and I think that’s him trying too hard. I think there is something to the fact that just like we all know, Ehlers also knows he’s on a much shorter leash than Kyle Connor. Ehlers is seventh in total minutes of ice time per game for the Jets, between bottom-six talents Mason Appleton and Morgan Barron.
There’s no way Ehlers doesn’t know.
Also, someone replied with a fun factoid to that Tweet:
Reddit Questions
Martin Necas is having a wonderful year in a contract season for the Carolina Hurricanes and is probably going to get paid. He definitely fits the Jets' mold, being a top-six winger who excels offensively but is a bit weaker defensively.
With both Sean Monahan and Tyler Toffoli as UFAs this summer, the Jets have two spots to fill in their top six. Cole Perfetti could easily be one of those pieces, but we’ll see.
I’m curious as to whether or not the Jets try to trade or extend Nik Ehlers next year. If they were to have Necas with this hypothetical roster and still have Ehlers, I would actually probably place him away from Ehlers. Necas is a great transitional player, and having him and Ehlers on separate lines would be wonderful.
Given Necas’ microstat profile similarities to Ehlers, he’d actually be a great candidate to potentially make 81-55 work. That would make the Jets' other top-six line 27-91-13, which I think would be interesting to see.
HOCKEY ANALYTICS EXPLAINED: When the Trade Market and Draft Market intersect and how to exploit them
Conspiracy theory time. I wrote this article on May 6, 2015. Then the Maple Leafs traded back almost every one of their picks. Kyle Dubas also unfollowed me on Twitter around the same time. Coincidence? I think not.
Getting serious, though...
There are two general types of research done on draft picks. One is looking at the projected value of a draft pick based on things like expected WAR or probability of getting an NHL regular. The other is looking at the trade value of draft picks in a market where you often see teams trading picks for picks.
What I found was that draft picks saw their actual value diminish from around picks #5 to somewhere between #24-40 faster than their trade value. This suggests that teams not desperate for a "likely NHLer" and wanting to maximize their expected value should trade back out of the first round if they don't have a top five-to-ten draft pick, depending on the year.
After around pick seventy-five, the diminishing value is about the same, but the trade value of picks is so low that trading back becomes advantageous again.
I like this type of research because it’s always interesting to compare the actual market relative to the theoretically optimal market.
FINANCE STUFF: Financial Literacy Part One of Four
The reason I incorporate financial and health/strength topics into these articles is for a few reasons. These are interests of mine, and I have spent a great deal of time learning about them because of it. Before I worked in hockey, I pursued a combined honors program with a focus on bioinformatics and analytical chemistry, but I also dabbled a lot in economics and finance on the side for fun because it was science-like electives.
The biggest reason is because this stuff matters, and it matters a lot more than hockey. Hockey is fun and a passion, but these are meaningful ways that can greatly improve your quantity and quality of life. My goal isn’t to tell people what to do, but rather to equip people with the tools and knowledge to be able to figure out what to do, or just make better choices.
Financial literacy matters. Even when studies include variables like socioeconomic demographics and formal education, financial literacy describes a huge variation in retirement wealth and similar outcomes.
Here in Canada, we are one of the more financially literate countries in the world, but it doesn’t say a lot. After all, Canadians also pay some of the highest fees for mutual funds in the world.
There was one 2019 study (Klapper, L., & Lusardi, A.) that used a very, very simple survey, and about two-thirds of Canadians passed. That survey was designed to be very basic too.
So, we’re going to go over each question in the survey over the next four weeks.
Question One: : Suppose you have some money. Is it safer to put your money into one business or investment, or to put your money into multiple businesses or investments?
I’ve heard more than one individual say that diversification is the only free lunch in finance.
Generally speaking, one has to increase their exposure to risk to increase their average expected returns on investment.
The exception is with diversification because the risk diversification removes is an uncompensated risk.
The stock market historically is a good investment. From 1916 to 2016, the average stock return was something like 7%. However, if you remove the top quarter of stocks, you not only get bad returns, but negative!
Stocks have a heavily positive skew:
This means it’s a few strong performers that pull the population up. The average stock underperforms the average of stocks.
So, by heavily diversifying, one creates a situation where they increase the odds that they do well on average.
This is in part why globally diversified index investing works so well. By having a piece of every business and getting the average return, you do better than the average person.
STRONGER STUFF: Suppliments and maybe avoid them
MASS (Monthly Applications in Strength Sport) Research Review is a monthly newsletter for super nerds in strength sports. It’s like The Five Hohl but for strength sports like powerlifting, bodybuilding, strongman, etc.
Anyways, one of their articles was recently released to the public on Stronger by Science, which means I can link it and talk about it.
It’s about supplements. We’re moving dangerously close to a lot of what my education was actually on in university, which I try to distance myself from here haha.
It should be noted that the focus of this write-up is on the US sports supplement world as that’s where most of the research the article is based upon was conducted, but there is a lot of overlap in different countries and their industries.
To summarize, the supplement industry is regulated and has oversight, but it’s a reactionary regulation as opposed to preventative like the pharmaceutical industry.
If someone were wishing to sell a drug commercially, they have red tape to cross with proving the safety, the effectiveness, and the contents of their product. The supplement industry tends to look into things typically after there are complaints, news, or research that comes out academically.
In addition, supplements tend to follow a three-step process to becoming commercially available. Many supplements are byproducts of other industries, like whey protein coming from the dairy industry. This is then sold to a manufacturer who will refine or create the supplement from the byproduct. The manufacturer then will sell that product to the supplement company, who may or may not have another middle partner deal with packaging, mixing of products, and labeling.
The combination of limited oversight and scrupulation, the process in getting the supplement to commercial availability, and financial incentives leads to some sketchy outcomes.
Research has shown that it’s very common for supplements to be tainted with materials not listed on the label. These tainted materials are often those that are banned or even illegal substances. It’s not uncommon that these tainted products are suspiciously tainted with products that align with marketed outcomes. For example, diuretics in weight loss supplements, hormones in testosterone boosters or protein powders, or methamphetamine in pre-workouts.
Not really discussed in the article, but supplements can also be lacking in what they are selling as well. Recently, turkesterone became a hot topic for supplements, and then it turned out that the commercially available supplements didn’t have the dose they were marketing. Protein powders may be “spiked” with things like glutamine, a protein amino that’s cheap but severely diminishes the efficacy of protein supplementation.
This is all to say that most supplements aren’t really worth it. There’s a risk that your supplement could be tainted or without the actual supplement you want to intake. It costs money and there’s a reasonable risk that you’re not getting what you want.
In addition, there’s not a lot of convincing research on most supplements anyways. Creatine is known to work better and more convincingly than any other supplement out there, and that has a huge number of non-responders and the impact is very small. The rest are either conditionally effective, like protein only if your diet is low in protein, or extremely uncertain, needing more research.
FINAL THOUGHTS
Thank you for joining us in this week's featurettes. We look forward to next week’s Monday Review, where we look at ongoing Jets story lines, prospects, and such.
If you've found this article or others enjoyable, please consider sharing it on your preferred social media platforms. I would really appreciate it if you follow us on Facebook and share this post to anyone you feel could benefit from our information!
Your support is greatly appreciated, as sharing helps us expand The Five Hohl and allows us to dedicate more time and resources to this project. It means a lot to me and I really appreciate your support. I’ve said “appreciate” a lot but it’s true.
Finally please subscribe. If you are financially able to, take a look at potentially upgrading to our paid tier. There you will recieve manually tracked microstatistics and analysis on said stats, the ability to comment on articles, and access to TFH’s archives.