This week we are returning to the playoffs one last time. I like to try and hold myself at least somewhat accountable for the recommendations we make here on Frozen Tool Forensics. We are returning to our Unlikely Playoff Heroes article from just before the playoffs. In that article I made a number of recommendations for players who had a better than average chance of outperforming their recent point paces and providing depth value for your playoff pool.

For context I am including a couple of quick excerpts from that article.

This all started back when I wrote the Wild West column. I saw a column from Rick Roos on playoff heroes. He looked at players who were performing better in the playoffs than in their previous regular season and attempted to figure out if it was predictive of future performance. I  decided to go the other way to look at if there was anything in the season data that we could use to predict which players might be poised to be this year's unlikely playoff hero for our playoff pools.

Generally speaking the criteria I found is:

" An unlikely playoff hero played 60+ games in the regular season. The player was averaging between 0.4 and 0.7 points per game, was shooting around two shots a game, with a least a shooting percentage of 9%, getting somewhere between 1:00 and 2:15 minutes on the power play, and skating between 13:00 and 17:00 minutes of total time on ice. The data points can fluctuate slightly depending on the situation, and players might miss one point by a little, but this is the general target. Perhaps the single most important factor though is that they play for a team that is likely to get at least 17 games during the playoffs."

The idea is that an unlikely hero is a player that is already doing a little something with the opportunity they are being given (to demonstrate some competence, value in the coach's eyes, etc.), but has room for growth. That growth could be either in deployment (increases in power play time, or total time on ice), or in favorable spikes of luck (team/personal shooting percentage, IPP etc.). A player who is already spiking a high shooting percentage or getting massive deployment isn't likely to see a sudden, relatively unanticipated increase in value

Using these criteria above, I cut the list of players participating in the playoffs down to a list of 17 players that had a better than average chance of improving during the playoffs. Given that the number of games was a very important factor I further provided five players, Tyler Seguin, Teuvo Teravainen, Taylor Hall, Seth Jarvis, and Mason Marchment who were on the teams odds-makers were giving them best shot at getting to the final four (Colorado, Edmonton, and Dallas in the West and Toronto, Carolina, and Boston in the East). Las Vegas and Toronto were also teams I advised reviewing.

So what is the postmortem?

Well for one odds-makers successfully predicted two of the four final four teams, and if we count my recommendations, we get three of the final four teams. Not terrible I suppose, but not exactly great. If we are confined to the five players highlighted because of those playoff-run predictions the results aren't great for the method. Only Jarvis proved more successful during the playoffs than the regular season.

If we look at the full list though the results are much more encouraging. Of the 17 players who met the initial criteria, 13 increased their point paces from the regular season (76 percent). If we apply a couple of filters from Rick's original article (at least ten games played, and an improvement of around 0.15 points per game) we are left with five.

So that means from our list of 17 original names, 75 percent saw improved performances in the postseason and five, or 30 percent of the names ended up meeting the criteria for an unlikely hero.

Overall this playoff season there were nine players who met the unlikely hero status – meaning over at least a ten game sample they improved their point pace by around 0.15 points per game, and totaled 0.65 points or better over the playoffs. Our list picked five of them, or 56 percent. Both in this case and in my original list of 17, I would have loved a higher hit rate, but overall it is pretty encouraging.

When drafting depth for the playoffs we are obviously not looking at top line players. Those are big names everyone knows and is going for. If we start big picture looking for depth, there were 148 forward players in the playoffs whose point pace going in was below 0.75 per game. That means technically the pool of unlikely heroes is 148 players. Of those 89 (or 60 percent) decreased their point paces in the postseason. Meaning only 59 or 40 percent saw an increase at all (compared to our 75 percent). That list of 148 contained all 9 unlikely heroes, but that is a hit rate of only six percent (compared to 33).

But not all of those players were realistic options, right? I mean most people aren't going to draft too heavily from a team's fourth line. So if we assume each of the 16 teams has five likely candidates (a combination of second and third line players, or maybe even a top line third wheel) that means the total pool of potential draft day targets is something like 80 players. That grouping should cover the entire nine heroes, but also gives you an 11 percent shot at grabbing one (versus the 33 percent hit rate above).

Big takeaway here, the hit rate using the criteria described above is certainly an improvement to throwing a dart at a middle six forward from a playoff team. Over the last several years the hit rate has been fairly consistent usually falling somewhere between 33 and 50 percent. At some point it might be worth doing another review of recent years and seeing if those data points can be refined somewhat, but that is an idea for another day.

For today, I wanted to close out with our Conn Smythe winner, Jonathan Marchessault. He was maybe a borderline candidate for inclusion in our original list as he has certainly had more productive seasons than this past one, so maybe wasn't as unlikely as others on the list. His Conn Smythe win does certainly indicate he was a hero though.

Over the course of the playoffs, he put up 25 points in 22 games (with 15 over his final 10) for a 93-point pace. His previous career high regular season pace was 80 points. During the playoffs he increased his time on ice and his shot rates. Playing on the top line with Jack Eichel certainly agreed with him. We should certainly stop and appreciate the performance he put up for Vegas.

That done, in terms of looking forward there are a lot of red flags that indicate this 93-point pace is not for real. His personal shooting percentage was up at 16 percent (up from his 11-12 percentage recent average). His team five-on-five shooting percentages was over 13, which is astronomical (his is typically closer to 8). His power-play IPP (points participation number) was 100 percent (typically 65ish), and at even strength it was 80 (typically 70-75), and 83 percent of his assists were secondary (typically 40ish). If he had been getting his usual results in those areas, we should have expected closer to a 60-point pace. That’s much more in line with his recent season numbers.

Marchessault didn't have Eichel to play with in all of those seasons so I suppose there is a bit of a reason for optimism next season that 60 points is the baseline, but he will probably command a lot of interest on draft day because of this performance. I might recommend passing and letting him be taken by someone who thinks he has truly found a new gear.

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