There are a couple of more playoff related topics that we are going to hit in this column, but they need to wait until the finals have wrapped up. In the meantime, let's return to the 2023-24 season as there are some additional ways we can look at the season, and some interesting reports we can run to help us. For this week we are going to look at a player's point production, but beyond that their consistency.

The Reports Page has a report titled "Most Consistent". In that report we get some basic scoring information about a player, and then a consistency percentage number. That number is basically the percent of games in which a player got a point. There is also a consistency percentage for goal scoring and assists. For the purposes of the article today we are going to focus on points.

First up, let's just take a quick look at who are the most consistent producers in the NHL.

It should come as no surprise that the folks with the highest point paces have points in the most games. Of course, Nikita Kucherov's 144 points are going to be spread out over more games than Anthony Cirelli's 45 points. In order for them to have similar consistency percentage numbers Kucherov would need to put up three plus points in every game Cirelli got a point. So since this is so clearly correlated to overall point production, why am I bothering to write about it?

Well, like with everything, nothing happens perfectly. We can account for someone's point pace and get an idea of who was more likely to spread out their production (basically get single points in more games), and who was more likely to pile on the points in a single game. Given the data set from the Most Consistent report I was able to compare point per game numbers and consistency percent numbers to get a basic trend line. It essentially allows us to say "given a player's points-per-game number, this is the percentage of games we would expect them to have points in." With that information and with a little rearranging of the report we have the top five players who were more consistent than expected. Broadly speaking it means that this group had more games with a point than expected and therefore were more likely than other players at their point pace to have single-point games versus multi-point games.

The second table is the players who were less consistent. Again, meaning they pointed in fewer games than expected given their point pace so had to have more multi-point games.

To illustrate what these tables are telling us, let’s use Cole Caufield and Gabe Vilardi as examples. Caufield played all 82 games and put up 65 points. That gives him a point pace of 0.79 points per game. Based on our expected numbers, we would anticipate someone with his point pace to have points in about 45 games. In order to get 65 points to fit in 45 games, we expect to see a bunch of multi-point games. If he only got one- and two-point games, we would have expected to see 20 multipoint games out of the 45. But the more 3+ point games he has, the fewer total multipoint games he will have if the number of one- and zero-point games remains constant. In actuality, Caufield had points in 53 games, eight more than expected. It also means he could have had at max 12 multi-point games. Because of a few three-point outings his actual number was 9 (so 44 one-point games).

Gabriel Vilardi on the other hand had very similar numbers, but was on the absolute other end of the spectrum in terms of consistency. He had many fewer games with points. As a quick illustration, he also had nine multi-point games but in 35 fewer games. Similar to Caufield, we would have expected him to point in about 45 games over an 82-game season. If we prorate his numbers to match an 82-game season, he would have pointed in 37 games, so in about eight fewer than expected. If this had carried on across the entire season, he would have had 16 multi-point games compared to Caufield's nine and 21 single point games versus Caufield's 44.

And now back to the tables. The first thing I noticed here was a lot of left wings in the first group and none in the second. That disparity holds true in that there are more left wings in the top five and top ten in the first measure, but by the time you get to top 20 the position numbers in both measures are pretty much the same. The disparity in the early ranks is likely just a coincidence and drafting a left wing probably isn't more likely to get you a player who points more consistently.

The second thing I noticed was age. This top five in the first group is generally younger than the second group. We might be able to rationalize something there with young players’ playing style being different than a veteran, but generally once you incorporate the top ten in each measure the average ages of the players seem to even back out. Ultimately I think this one is mostly noise as well.

Finally, I wanted to check and see if this was something that players repeated year over year. I took the same report for 2022-23 and did the same calculations. In the second group, basically everyone was in a much different place in this ranking during the 2022-23 season. Almost everyone was in the positive percentages and Wyatt Johnston was at almost five percent. So far repeatability isn't looking great.

For the first group, Cole Caufield was the most consistent player in 2023-24 at almost nine and a half percent more likely to score a point in a game than his point pace would indicate. In 2022-23 he was about a half a percentage point below average. It was a swing of about 10 percent. Again, very not consistent.

Matias Maccelli, Jake Guentzel, and Nicholas Robertson on the other hand were all much closer. Maccelli, who ranked second at 7.9 ranked 19th in 2022-23 with a 5.1. That seems incredibly consistent compared to the rest of the named players here. He actually falls into a fairly typical range of players though. More than half of all players in the set (360 out of 611) saw a three percent change or smaller in this stat. So on the one hand kind of repeatable? The range is only 10 percent in either direction, and most players see swings of less than 3 percent. Ultimately though I think that is just the nature of thing we are measuring and not necessarily an indication of an individual player's impact on the data.

So what is the takeaway here? Best case scenario was that it might give managers some ability to design a team that was more consistent or more boom and bust depending on what they wanted, but unfortunately it doesn't really look like we can do that. Ultimately there doesn't appear to be a huge amount of correlation year or year, or a lot of proactive information we can use here moving forward for next season. This measure and stat seem like a good way to retroactively describe what happened during a player's season and put some actual numbers behind the differing experience that managers had when rostering Caufield versus Vilardi.

That is all for this week. Do your part to support organizations working to make hockey for everyone.

Black Girl Hockey Club

Hockey is For Everyone

Athletes for Hope

Premier Hockey Federation

Editor's note: Frozen Tools Forensics will return on July 19.