Steven Stamkos is the best shooter of the salary cap era. There. I said it. I know right away some Washington Capitals fans will be screaming about Alex Ovechkin. And Toronto Maple Leafs fans screaming about Auston Matthews. And Winnipeg Jets fans screaming about Patrick Laine. I hear you.
But if we look at all the evidence, much of it points to Stamkos being the best shooter since the salary cap was instituted in the NHL. Ovechkin is without a doubt the best goal scorer in that time period but he’s done that with a high volume of shots, not with elite accuracy. Matthews and Laine are putting themselves in the conversation, but have only been in the league for a few years. Four or five years down the line, if they’ve kept up their shooting prowess over a longer span, this will be an interesting conversation to revisit.
In starting my research for this idea, I explored different ways of measuring how good of a shooter a player is. Each metric has strengths and flaws. But I think in the end, the data supports the conclusion that Stamkos is the best shooter of this era. I’d like to thank my editor, loserpoints, as well as the guys at Evolving-Wild.com for their help and guidance. Without them, this article would not have been possible.
In this article, I’ll focus on Ovechkin, Matthews, Laine, Brad Marchand, Connor McDavid, and Sidney Crosby as the other players to use as comparisons to Stamkos. The first three are obvious. They’re Stamkos’ closest rivals for the marksmanship crown. Marchand is less obvious but he’s a surprisingly good shooter for a rat. McDavid and Crosby are two of the greatest players in the game and serve as nice benchmarks for just how good the other shooters are.
Shooting Percentage
Shooting percentage is the easiest and most accessible stat to use. The flaws in it are that it doesn’t account for shot difficulty in any way or for power play vs. even strength. A player who always shoots from right around the net is naturally going to have a higher shooting percentage because those shots are more dangerous and thus, easier to convert. A shooter that shoots from the walls or the point (like defensemen) are generally going to have a lower shooting percentage because those are easier saves for a goaltender to make.
We also have to take sample size into account. Shooting percentage is prone to fluctuations in small samples so players who haven’t taken many shots could have inflated percentages. Maybe they were hot. Maybe they were lucky. Maybe they were just exceptional at taking advantage of the few shots they got even though they were a third or fourth liner.
At a high level, we can go to NHL.com and use the player stats page to get accumulated stats for all players since the 2005-06 season. First, let’s constrain for players that have played at least 200 games so that we focus only on players with significant NHL careers.
Right off the bat, Stamkos came in at 5th overall with a 16.9% shooting percentage. He’s led by Alex Tanguay, Paul Byron, Brayden Point, and Andre Brunette. Let’s compare these four , Stamkos, and our benchmark group in some key stats to consider here.
Shooting Percentage
| Player | Games Played | Goals | Shots | Shooting Percentage | Shots per Game | TOI M | TOI S | Shots Per 60 |
|---|---|---|---|---|---|---|---|---|
| Alex Tanguay | 709 | 175 | 967 | 18.10% | 1.36 | 17 | 51 | 4.58 |
| Paul Byron | 419 | 85 | 482 | 17.60% | 1.15 | 14 | 29 | 4.77 |
| Brayden Point | 229 | 91 | 530 | 17.20% | 2.31 | 18 | 38 | 7.45 |
| Andre Brunette | 568 | 147 | 866 | 17.00% | 1.52 | 16 | 4 | 5.69 |
| Steven Stamkos | 746 | 393 | 2322 | 16.90% | 3.11 | 19 | 27 | 9.60 |
| Alexander Ovechkin | 1084 | 658 | 5234 | 12.60% | 4.83 | 20 | 58 | 13.82 |
| Auston Matthews | 212 | 111 | 717 | 15.50% | 3.38 | 18 | 4 | 11.23 |
| Patrick Laine | 237 | 110 | 690 | 15.90% | 2.91 | 17 | 11 | 10.17 |
| Brad Marchand | 681 | 262 | 1657 | 15.80% | 2.43 | 17 | 29 | 8.35 |
| Connor McDavid | 287 | 128 | 870 | 14.70% | 3.03 | 21 | 22 | 8.51 |
| Sidney Crosby | 943 | 446 | 3063 | 14.60% | 3.25 | 20 | 53 | 9.33 |
The first thing that sticks out about the four players ahead of Stamkos is that they shoot at a much lower pace than Stamkos. Obviously, he has more ice time than others, especially Byron and Brunette, but he’s shooting more often when he’s on the ice than them. The right hand column shows shots on goal per 60 minutes of ice time for each player. You can also see that Ovechkin, Matthews, and Laine, are all shooting at a higher volume than Stamkos.
Brayden Point has a stronger argument than Byron and Brunette. However, he’s on the lower end of the games played scale along with Matthews, Laine, and McDavid. He was powered by a very high shooting percentage in 2018-19 as he jumped from 14.8% and 14.7% to 21.5%. Based on what we know about the fluctuations in shooting percentage, his will likely come down and he is likely to end up out of the conversation pretty quickly.
Alex Tanguay has a good case here as he played a lot of games and has a higher shooting percentage. But you have to ask the question of if he had been shooting more, would he have been scoring goals at the same rate? And I have my doubts. Tanguay was a consistent point producer in the top six, but he was much more of a playmaker on the ice as evidenced by his 366 assists compared to 175 goals in the salary cap era. He finished his career with 283 goals and 580 assists in 1,088 games. To me, this says that he was very selective with his shots and was taking the passing option more often. But when he did take a shot, it was a dangerous one, which would naturally lead to a better shooting percentage.
As a stat, raw shooting percentage has some limitations. It includes all situations and doesn’t account for shot difficulty at all. For Stamkos, Ovechkin, and Laine, their goal scoring on the power play is a big part of their games. Matthews is good at getting to the front of the net at even strength and on the power play. So that leads us into our next stat that will give us a better idea of how they shot vs how they were expected to shoot based on the difficulty of the shot.
Goals Scored vs Expected Goals Scored
Expected goals models help us to know what we would expect a player’s shooting percentage to be. Thousands of games have been tracked to give us shot locations and results from all over the ice. By taking this large sample size, analysts create models to map the average shooting percentage of a shot taken from a certain spot on the ice. In general, shooting percentage is high around the front of the net and goes down the further away from the net you go.
Unlike scoring chance calculations, these models don’t map out general zones to assign them shooting percentages. Instead, every point on the ice surface has it’s own assigned shooting percentage based on past results. This can also be broken down by even strength, power play, and shorthanded, as well as the other variations of players being on the ice.
This modeling allows us to look at how many goals a player should have scored if they were an average shooter with the shots that they took. If a player scores goals at a higher rate than that, we can surmise the player is a better than average shooter. Conversely, the player may not be a great shooter if they are scoring below the average.
The flaw in this approach though is that models rely heavily on shot locations and shot types to assess the danger of the shot. Some models also include some estimations about rebounds and shots off the rush based on proximity in time and space to events just prior to a shot. Missing pieces from the models include the locations of the players during the shot as in the case of screens and the passes made just prior to the shot. And of course, no model could ever account for just plain old pure luck.
The following stats are taken from Evolving-Hockey.com. Unfortunately, play-by-play only goes back through the 2007-08 season, so we do not have information from the 2005-06 or 2006-07 seasons.
Goals vs Expected Goals
| Player | GP | TOI | EV G60 | EV ixG60 | Difference | PP TOI | PP PPG60 | PP ixG60 | Difference |
|---|---|---|---|---|---|---|---|---|---|
| Steven Stamkos | 746 | 10921.45 | 1.2 | 0.75 | 0.45 | 2742.64 | 3.11 | 1.88 | 1.23 |
| Patrik Laine | 237 | 3237.28 | 1.09 | 0.65 | 0.44 | 721.07 | 3.41 | 1.93 | 1.48 |
| Nikita Kucherov | 447 | 6430.61 | 1.12 | 0.76 | 0.36 | 1303.97 | 2.44 | 1.78 | 0.66 |
| Auston Matthews | 212 | 3213.57 | 1.55 | 1.21 | 0.34 | 507.2 | 3.08 | 2.35 | 0.73 |
| Ilya Kovalchuk | 493 | 7850.8 | 1.02 | 0.69 | 0.33 | 2382.93 | 1.59 | 1.29 | 0.3 |
| Alex Semin | 519 | 7260.04 | 1.11 | 0.81 | 0.3 | 1648.42 | 1.6 | 1.55 | 0.05 |
| David Pastrnak | 320 | 4507.2 | 1.09 | 0.8 | 0.29 | 766.77 | 3.29 | 2.27 | 1.02 |
| Guillaume Latendresse | 261 | 3077.59 | 1.03 | 0.74 | 0.29 | 445.86 | 1.75 | 1.89 | -0.14 |
| Jake Guentzel | 204 | 3052.97 | 1.18 | 0.89 | 0.29 | 371.67 | 2.26 | 1.71 | 0.55 |
| Andreas Athanasiou | 248 | 3049.75 | 1.2 | 0.92 | 0.28 | 361.13 | 0.83 | 1.27 | -0.44 |
| Brad Marchand | 681 | 9362.18 | 1.03 | 0.77 | 0.26 | 1173.05 | 2.4 | 1.49 | 0.91 |
| Sidney Crosby | 782 | 12280.03 | 1.16 | 0.9 | 0.26 | 3263.99 | 1.97 | 1.6 | 0.37 |
| Alex Ovechkin | 920 | 14475.2 | 1.26 | 1.01 | 0.25 | 4112.48 | 3.06 | 2.23 | 0.83 |
| Vladimir Tarasenko | 497 | 7095.62 | 1.22 | 0.97 | 0.25 | 1331.37 | 2.48 | 1.76 | 0.72 |
| Rick Nash | 775 | 10830.5 | 1.14 | 0.91 | 0.23 | 2282.4 | 1.68 | 1.82 | -0.14 |
| Sean Monahan | 471 | 6782.46 | 1 | 0.78 | 0.22 | 1395.12 | 2.02 | 2.28 | -0.26 |
| Brayden Point | 229 | 3358.33 | 0.98 | 0.87 | 0.11 | 612.22 | 2.84 | 1.75 | 1.09 |
| Mikko Rantanen | 238 | 3531.18 | 0.7 | 0.52 | 0.18 | 787.39 | 2.44 | 1.18 | 1.26 |
| Elias Pettersson | 71 | 1008.9 | 0.89 | 0.62 | 0.27 | 252.4 | 2.14 | 0.94 | 1.2 |
| Patrick Eaves | 502 | 5283.17 | 0.67 | 0.75 | -0.08 | 606.83 | 2.97 | 1.99 | 0.98 |
This table shows quite a few more players. It also splits the stats into even strength and power play. Most of the players are on this list either because we included them in our original group for analysis or because they had a high goals per 60 rate. I added a few more at the end because of their excellence on the power play and included Elias Pettersson because he’s an intriguing young player.
First, we have the even strength stats: Time On Ice, Goals Per 60, Individual Expected Goals Per 60, and Difference. By using the rate stats, we’ve evened the playing field as far as ice time is concerned. It doesn’t matter how much the player has played, it matters what they’ve done with their opportunity. Like before, I limited the players to those with at least 200 games played, with the exception of Pettersson who I threw in for fun after seeing his power play stats.
Stamkos tops the list in exceeding his expected goals, doing so by 0.45. That means that he scores 0.45 goals more per 60 minutes on the ice than we would expect. Laine is right behind him at 0.44. Nikita Kucherov also makes an appearance exceeding his expected goals by 0.36. And next on the list is Auston Matthews. You’ll notice that no one, not even Alex Ovechkin, exceeds Matthews in goals per 60. So how does he do it? Well, he gets to the front of the net. A lot.
While not included in this table, Matthews gets off 15.73 shots on goal per 60 (iSF60). Stamkos by comparison has 16.17 iSF60. By comparing those two numbers and their ixG60, you can see that Stamkos shoots slightly more than Matthews, but does so from further away from the net. Ovechkin, by comparison, takes 20.27 shots per 60 minutes. Ovi is the only other player on our evaluation list that exceeds 1 expected goal per hour.
The table also includes power play stats. Stamkos exceeds Ovechkin in power play goals per 60 with 3.11 to his 3.06. But Laine beats out both with 3.41 PPG60. David Pastrnak is also a significant data point with his 3.29 PPG60. Austin Matthews joins the group of players over 3 with a 3.08.
Once again when looking at what is expected of them, Matthews leads the way as he gets around the net on the power play just as he does at even strength for his shots. Stamkos, Ovechkin, and Laine are known for their one timers, which leads to a slightly lower ixG60 because of the distance from which they take those shots. Laine leads the way in exceeding his ixG60 by 1.48 goals per 60. Stamkos is right up there with him at 1.23. Mikko Rantanen made it onto the list because he also exceeds his ixG60 by 1.26, but his expected goals rate is much lower at 1.18. Point, Pettersson, and Patrick Eaves also made it onto this list because of their ability to exceed their expected goal rates.
In this category, we can call Stamkos and Laine pretty much equal. Laine gets a slight edge for outperforming more on the power play than Stamkos does. However, Stamkos has three times the sample size that Laine has. Laine has only been at this for three seasons, while Stamkos has been in the league for 11 seasons. We’ll need to keep an eye on these numbers for Laine going forward to see if he is able to sustain this success in the long term as Stamkos has.
Goals Above Average Per Fenwick
Fenwick is a measure of all unblocked shots including goals, shots on goal, and missed shots. In short, it ignores any shots that are blocked. Goals Above Average is a stat used more with goaltenders to show how they impact the odds of a goal being scored. However, it also has a shooter component. A skater with a positive GAA Per Fenwick is increasing the league average shooting percentage, while a player with negative GAA Per Fenwick is lowering the league average shooting percentage. One of the advantages of this model is it accounts for other factors such as the expected value of the shot, the goalie, strength, score, and player’s position, among other factors.
For this metric, I selected all available data from Evolving-Hockey.com. You can find the full tool under the “Misc” tab listed as Shooting GAR Tables. I then constrained for a minimum iFF (Individual Fenwick For) of 750. This removes most of the sub-200 game players just as we’ve been doing for the rest of our analysis. First, doing a quick sort by Goals Above Average (GAA), Stamkos and Ovechkin jump right to the top of the list and then it’s a big drop off from those two to Sidney Crosby, Brad Marchand, and Patrick Kane.
What’s a bit more amazing here is that Stamkos leads Ovechkin by 24 GAA in a period where Stamkos wasn’t playing for a season, missed half of another season, and missed all but 17 games in a third season. In that span, Ovechkin only missed 29 games while Stamkos has missed 122 games, plus the 82 games from the extra season Ovechkin has in the data set.
Ovechkin also has a big advantage in iFF with 6,300 while Stamkos only has 3,242. As a comparison, Crosby is at 3,299 iFF and yet is only at 89.6 GAA. That means Stamkos has accumulated GAA at an over 40% higher rate than Crosby. It also means that he has positively impacted the league average shooting percentage more than any other player in the league.
The following chart is a histogram showing the frequency with which players have reached a certain GAA total. The position of Stamkos and Ovechkin here shows just how much they’ve outpaced the competition. Even Crosby hangs far back of the leaders while being clearly ahead of the rest of the pack in third place.
However, to be fair to the players that haven’t been in the league as long as Stamkos and Ovechkin, we need to look at the GAA Per Fenwick since this is a rate stat, instead of a counting stat. Though so far, I think we have to lean toward Stamkos because of how far ahead of the field he is.
Sorting by GAA Per Fenwick, we still end up with Stamkos at the top with 0.0490. Laine comes up second at 0.0455. The next step down goes to Brad Marchand at 0.0344 and Kucherov at 0.03357. A somewhat surprising name to see up here is Brett Connolly at 0.031.
Goals Above Average Per Fenwick
| Player | TOI | GAA Per Fenwick | iFF | GAA | GAR |
|---|---|---|---|---|---|
| Steven Stamkos | 14496.5 | 0.04907 | 3242 | 159.1 | 180.8 |
| Patrik Laine | 4064.5 | 0.04554 | 975 | 44.4 | 50.9 |
| Brad Marchand | 11893.2 | 0.03435 | 2186 | 75.1 | 89.7 |
| Nikita Kucherov | 7959.1 | 0.03357 | 1823 | 61.2 | 73.6 |
| Brett Connolly | 5171.1 | 0.03084 | 788 | 24.3 | 29.8 |
| Artemi Panarin | 6276.2 | 0.03074 | 1132 | 34.8 | 42.3 |
| Alex Tanguay | 9883.5 | 0.03016 | 1008 | 30.4 | 37.1 |
| Leon Draisaitl | 6650.8 | 0.03015 | 1048 | 31.6 | 38.4 |
| David Pastrnak | 5412.4 | 0.02927 | 1288 | 37.7 | 46.3 |
| Ilya Kovalchuk | 10716.2 | 0.0292 | 2281 | 66.6 | 81.9 |
| Erik Haula | 4867.6 | 0.02813 | 782 | 22 | 27.2 |
| Aleksander Barkov | 8204.5 | 0.02809 | 1253 | 35.2 | 43.6 |
| Jiri Hudler | 9482.4 | 0.02727 | 1320 | 36 | 44.9 |
| Sidney Crosby | 16380.9 | 0.02716 | 3299 | 89.6 | 111.6 |
| Auston Matthews | 3831.2 | 0.02608 | 997 | 26 | 32.7 |
| Alex Ovechkin | 19195.3 | 0.02125 | 6300 | 133.9 | 176.1 |
| Patrick Kane | 18022.3 | 0.01959 | 3792 | 74.3 | 99.8 |
| Evgeni Malkin | 15500 | 0.01958 | 3463 | 67.8 | 91.2 |
| Phil Kessel | 17166.3 | 0.01789 | 4053 | 72.5 | 99.7 |
| Connor McDavid | 6126.1 | 0.01767 | 1126 | 19.9 | 27.5 |
This list is the top players by GAA Per Fenwick and also some other top players that have accumulated a large amount of GAA in their career. While Laine is up there, Stamkos is edging him out. The Lightning captain has also kept it up over three times the amount of unblocked shot attempts.
If we look at both parts of the calculation for GAA, GAA per unblocked shots and unblocked shots, we can again see just how much Stamkos sticks out on the plot. In this case, unblocked shots is converted to a rate stat to level the playing field among the players in terms of career length. The circles for each player are sized by total GAA.
Again, we see that Stamkos is an outlier in terms of shooting skill and Ovechkin an outlier in terms of shot volume. The smaller gray dot next to Stamkos is Laine.
Digging further into the numbers, Laine started off hot as a rookie posting the best GAA Per Fenwick season in the data set, came back down toward Stamkos’ level in his second season, but really dropped off during his third season. Perhaps it will just be a momentary blip and we’ll laugh when we look back on this in a few years. But for now, the question of if he’ll bounce back to his previous form or something close to it is legitimate. Because if he doesn’t, he’ll fall back from Stamkos quickly.
In terms of career trajectory, Stamkos started off a little slower, jumped up in his second season, fell back towards the middle in his third season, and then exploded when he scored 60 goals. But perhaps what’s most impressive is that he had his third highest GAA and second highest GAA Per Fenwick season in 2018-19. So if we’re looking for him to slow down as a shooter, he hasn’t started yet.
Among players with at least 200 iFF in a season, Stamkos has three of the 25 best GAA Per Fenwick performances. Laine’s rookie season is at the top of the list and Stamkos’ 60 goal season is second. Laine has two spots and is the only other player to show in the top 25 more than once. In terms of total GAA, Stamkos has 5 of the top 25 GAA seasons including the top two spots. Ovechkin has three seasons in the top 25. Laine and Crosby have two each.
Conclusions
After looking at all the evidence, the debate comes down to Stamkos and Laine. Both are near or at the top in all of the metrics we’ve examined. But to me, the difference is that Stamkos has maintained this success over a much longer period of time. I know that’s not quite fair to Laine, but it also wouldn’t be fair to Stamkos to discount his prolonged success. Stamkos has played in 746 career NHL games over 11 seasons. Laine has played just 237 games over three seasons and is just beginning his journey.
Ovechkin will assuredly go down as the best goalscorer of the Salary Cap Era, and maybe of all time. His eight Rocket Richard trophies say everything that needs to be said on that topic.
But when we’re talking about the best shooter? It has to be Steven Stamkos. Laine may make a claim for the title down the road, but for now, Stamkos wears the crown.
