Head-to-head leagues: Most consistent players and why it matters

Auston Matthews is a steady, consistent building block for your fantasy hockey roster. Julian Avram/Icon Sportswire

Does it matter when your fantasy players get their points?

In a rotisserie fantasy league, it all comes out in the wash. You can make the same argument for a season-long points format.

But in head-to-head, it matters a lot. If you have players that disappear some weeks, it can have a major impact on how you do in the standings.

In an ideal world, your players' fantasy points would be spread out as evenly as possible across all 82 games.

That's where consistency comes in.

Do you want to know who was the most consistent fantasy player per game from March 1, 2022, to Nov. 15, 2022, across parts of two seasons?

It was Kyle Burroughs of the Vancouver Canucks if you are OK with a smaller sample of just 16 games played; it was Alec Martinez of the Vegas Golden Knights if you prefer a larger sample of 32 games.

How about which player with at least 150 games during the past two years has the most outliers among their fantasy performances?

It was Pat Maroon of the Tampa Bay Lightning, who had 17 games over the past two seasons that would count as statistical outliers.

What about most consistency per game in the first 10 games of last season?

Well, that would be Devon Toews of the Colorado Avalanche, who posted 2.2, 3.5, 2.1, 1.8, 1.7, 2.1, 2.0, 3.6, 3.3 and 2.1 points over those 10 games.

Define consistency

Let's talk about how we are measuring consistency here, as there are many ways to do so.

Standard of deviation is the easiest, as it is a measure of the spread of a player's fantasy production when compared to their average (mean). A low standard of deviation means most of a players fantasy outings are close to what they produce most of the time. A high standard of deviation would indicate the player has wild swings in how many fantasy points they might produce from game to game.

But that is not good enough to compare Player A to Player B. As a standard of deviation will be relative to the player's average production. Coefficient of variation offers a standardization by comparing that deviation to the mean to give a better comparison, which can be displayed as a percentage.

From the examples above, Burroughs and Martinez have coefficients of variation (CV) of 42.6 per cent and 48.4 per cent, respectively, between those two dates. Toews's CV across those 10 games at the start of last season was 29.8 per cent.

CV still isn't perfect, as it is still measuring data spread relative to mean, and being consistently irrelevant for fantasy is still consistent. But we can combine what we know about fantasy players to select some to focus on that manage to produce the lowest CV over the past few seasons.

The data

This look at consistency is pulled together using game logs from the 2020-21, 2021-22 and 2022-23 regular seasons. In generating a CV for each player, I also generated how many games they played in which their fantasy points would count as a statistical outlier -- as referenced above with Maroon.

That is when the result is so far away from the mean for the player that it is, well, an outlier. We could go through an exercise in which we remove outliers from each player's dataset and generate a CV based on just the games that are not statistical outliers. But I thought it better to have the CV separate and then note how many outliers they had. This is especially true because we are using fantasy points as our guiding light for consistency. Because most skaters are below 3.0 fantasy points per game, it's actually not very likely at all (if almost impossible) for a skater to have a low -- or in fantasy parlance, "bad" -- statistical outlier.

Because the outliers for skaters are a positive thing, they "popped off" for a game, knowing about the outliers is actually a positive bonus here.

Goalies, because they can score negative fantasy points, are a different story, of course.

Finally, I chose an admittedly arbitrary cutoff of 100 minimum games before considering a player for the final table of consistency.

Starting your own league is easier than ever! Set your league size, select your keepers, and invite your friends to start playing. Sign up for free today be the commish!

Consistency Heroes

As one might expect, defenders take the cake on this. Of the top 100 players with the lowest coefficient of variation over the past three seasons, 85 of them play on the blue line.

Moritz Seider, D, Detroit Red Wings (164 games played, 2.24 fantasy points per game, 53.8% coefficient of variation, four outliers): It's not a surprise he's on the list, but it might be a surprise that he tops it. Seider has the advantage, if you can call it that, of only playing two of the three seasons we are examining, which is less time to have the occasional rough outing. But you can set your watch to Seider's production and there is no reason to think that will change.

Andrew Peeke, D, Columbus Blue Jackets (173 GP, 1.66 FPPG, 54.9% CV, two outliers): No, don't draft Peeke. At least not in your average-sized fantasy hockey league. I want to include him here to point out that he's second overall for CV percentage, but also to underscore how consistent does not equate with fantasy relevance. In Peeke's case, he just doesn't quite make the cut for fantasy consideration. And with Ivan Provorov and Damon Severson invading the Blue Jackets blue line, there's no reason to think Peeke will find the ice time to suddenly get into the picture. But good on him for being the model of consistency while blocking those shots.

Darnell Nurse, D, Edmonton Oilers (209 GP, 2.21 FPPG, 55.1% CV, one outlier): With the emergence of Evan Bouchard in the playoffs and pomp that went along with Matthias Ekholm's arrival, don't forget about Nurse. He did his job last season with only one power-play point, so Bouchard coming on strong should have no impact on Nurse's value.

Auston Matthews, C, Toronto Maple Leafs (199 GP, 3.23 FPPG, 59.1% CV, two outliers): The top forward for consistency over the past three seasons is among the top two players for fantasy. Big surprise. In fact, Connor McDavid (218 GP, 3.47 FPPG, 62.4% CV, two outliers) ranks second in this evaluation. So, maybe we aren't learning a whole lot when it comes to forwards. But, as I'll get into in a minute below, a high CV isn't necessarily a bad thing, either, when it comes to fantasy. But have a look at the other forwards in the top 100 for low CV.

The other forwards in the top 100 for coefficient of variation and their fantasy points per game:

How do you feel about them? Are they, for the most part, the most trusted fantasy forwards you can draft?

Look at some of the forwards with the highest CVs and strong FPPG showings:

How do you feel about these elite producers? Do you trust them on your fantasy squad? Are they as worthy of a high investment as the list above?

Maybe all this statistical evaluation really brings us to a quantifier of fantasy trust, more than it does fantasy consistency.

The rest of the top 10 in CV and their fantasy points per game:

Ready to play? Sign up for free today!

Consistency Zeroes

But wait... outliers in the context of skaters for fantasy are actually a good thing. Skaters don't score negative value, so outliers are a massive boost to their overall productivity. So does having a higher coefficient of variation actually mean the players are a better fantasy play because they have a chance of explosive production that increases the spread of their data?

The answer is, sort of. See, none of these factors can be viewed in isolation. It's all just more information to help you evaluate how you feel about certain players. Let's look at an example.

Tage Thompson has the highest CV among all skaters that managed to average at least 1.7 FPPG across the past three seasons. So his game log of fantasy points across 194 games has the widest variation. But outliers go a long way to pushing that spread, thus making his standard deviation and coefficient of variation higher than other players.

Let's remember one key thing though: Skaters don't score negative fantasy points. So the "outliers" to Thompson's dataset are games in which he scored 10.9, 13.4, 8.2, 7.2 and 7.7 fantasy points. That is most certainly not a bad thing for your fantasy team. Inconsistency in this case is explosive production.

So what's the minimum number of outliers before we should consider a player as "inconsistent" for fantasy in this context? Is it two? Maybe just one explosive game over three seasons isn't enough to warrant concern?

I'm going to argue that it's zero. So to consider a player in this negative connotation of inconsistency, we will only select from those with zero outlier games. As in, over the past three years, they haven't "popped off" for a night once.

Valeri Nichushkin, W, Colorado Avalanche (170 games played, 1.76 fantasy points per game, 86.4% coefficient of variation, zero outliers): Here is where inconsistency and no positive outliers meet. Nichushkin has been giving much better per-game results the past two seasons and some of this inconsistency probably stems from before he finally broke out in 2021-22. He's missed a lot of time in his career and the Avs don't have the top six that they did two seasons ago. Plus we have the incident in the playoffs, of which clarity doesn't appear to be forthcoming. But he's back with the team and is coming at drafts at a steep discount. So maybe mercurial output is acceptable given the price of adding him to your squad.

So who is the CEO?

That stands for consistent every outing.

Now that we've touched on the concept of outliers actually being a positive thing, you may still be wondering: Who is the most consistent fantasy player, regardless of the fact that outliers are actually good in our world?

Steven Stamkos is the answer. In 200 games over the past three seasons, he's averaged 2.43 FPPG with 73.5% coefficient of variation and not one outlier in his game log.

Alex DeBrincat follows close behind with 216 games, 2.24 FPPG and also zero outliers.

But again, as we've discussed, the more outliers the merrier in this context.