Thoughts on the new Enhanced Statistics site on NHL.com

I wasn’t actually planning on writing anything formal about the new enhanced hockey stats on NHL.com but this post over at Jewels From The Crown was kind of the last straw.

Before getting into that article let me say a few things. Despite the fact that I run a popular hockey stats site I really wanted to see the NHL do a good job on their advanced hockey stats site. I honestly don’t see them as a competitor nor do I really care if they are because I make no money off the site and my interest lies as much in analysis and research as it does in producing and running a website. I also see the NHL.com site being more geared to the average, more casual user while my site is geared more towards the hard core user and researcher. I love hockey, I love hockey statistics and hockey analytics, and I really would have loved to see the NHL do this right to bring this to a wider audience than I, or any of the other hockey stats sites, ever could. While I still have that hope my thoughts on this first attempt is that is a very poor effort that could have gone much better.

So, now, what set me on this bit of a tirade. Well, the post at Jewels From the Crown featured an interview with Chris Foster, NHL Director of Digital Business Development, and Gary Bettman. In it Sheng Peng asked about what ‘exclusive’ stats that NHL.com offers over other sites such as mine. This was a portion of the answer.

Foster: I’ve got to double check. I’m not sure. There’s zone starts, I think those are completely brand-new. And the level of depth that we’re doing with primary and secondary assists. I don’t think anybody’s going to have that much detail. That first batch—shot attempts and unblocked shot attempts–there’s a lot of that. It’s that second batch of stats—primary assists and penalties drawn over time—those are the ones that will be more unique to the site. They may be out there but not to the level of depth that will be on NHL.com.

Zone starts?  Brand-new? Really? Zone starts have been around for years. Primary assists are exclusive to NHL.com? Really? I’ve had them on my site for years too. I even go a step further and look at primary points (goals + primary assists). Penalties drawn has been around elsewhere too. I’ll give the NHL the benefit of doubt and believe that they are actually oblivious to what else is being done out there because otherwise they are outright misleading and belittling the hard work that I and many others have done previously. Looking at their enhanced statistics site it is clear that they haven’t really put much thought into this whole project or reached out to the analytics community because they are tons of things that I think many would suggest they do differently. Here are a few examples:

  1. SAT and USAT are short for Shot ATtempts and Unblocked Shot ATtempts otherwise known as Corsi and Fenwick respectively. I am OK with the name change but for the NHL’s target audience it is absolutely unnecessary to use both. Even myself as an analytics person at times wonder why we have both Fenwick and Corsi. They are extremely highly correlated and the benefits of one over the other is generally very minimal. For the casual user it is completely unnecessary to burden them with these two separate stats. It would have been far better to simply use shot attempts (Corsi) and leave out the unblocked variety. Shot attempts are simple, straight forward, and easy to understand what they are.
  2. The Skater Shooting/Time on Ice stats have both /20 and /60 statistics which is redundant and pointless. One is just 3 times the other. Why one would see the need to present both side by side on the same page is beyond me. Let’s present a stat. Ok, now lets multiply it by three and present that too. Who thinks like that? Really? Who? Furthermore, when I started my site I used /20 stats because I figured a good player plays about 20 minutes per game. Other sites used /60 because a game is 60 minutes long and it tells how that caliber of player would produce in a full game. Both have merit but for the purpose of consistency across I have converted all my stats to /60. Had they reached out to me I’d have told them this and they may very well have done the smart thing and just present /60 stats.
  3. Having a stats site and not being able to filter based on games played or time on ice it practically useless. When I sort by SAT% I want to see who the best players are who play regular or semi-regular shifts. Instead the top of the list is dominated by AHL call-ups with one or two games that nobody has heard of and nobody cares about. Why do this?

There are numerous other smaller mistakes as well (see Eric Tulsky’s twitter time line for a few of them). It’s a shame really because I was hoping for and expecting for a whole lot better. I applaud the NHL for hopping on the ‘enhanced statistics’ bandwagon but what they released today screams of a poorly thought out beta release of a product developed by a group of amateurs, a long way from a major new product release from a multi-billion dollar organization (NHL) backed up by another multi-billion dollar organization (SAP) which they promoted it as being.

I really do hope that the NHL gets their act together and makes it work as I think it will be good for everyone. The NHL, the casual fan, and those in the hockey analytics community. We all benefit when the NHL does things well. We are, at the core, all hockey fans. My hope with this post is that it inspires the NHL to spend more time reaching out to the people that have been doing this for years. We have years of experience, knowledge and expertise that would have helped avoid many of the basic and senseless missteps we see today. If you are with the NHL and are reading this I want you to know that I more than willing to share my experience and I am certian most everyone in the hockey analytics community would as well. You just have to ask. My e-mail is david@hockeyanalysis.com.

This article has 5 Comments

  1. love the post but coming from an average/beginner analytics user, having both Corsi and Fenwick available is welcomed. It is fun and interesting to be able to compare them. That’s it, no rant here. It’s just fun to have them both there. Please don’t speak for us, we’re just trying to get as good at this as we can.

    1. That’s fair. I’ll only say that having compared them a lot the differences are minimal and really not very important or valuable. I personally think you’d get more value looking at other statistics that matter more.

  2. Why would anyone at the NHL care that you are “more than willing to share my experience” after you just called them and their product a “poorly thought out beta release of a product developed by a group of amateurs”? Just curious.

    1. I’ll leave that for them to decide. I think my critique is more than valid for all the reasons I specified. It will be up to them to figure out how to rectify the problems that led up to them releasing what I, and most everyone in the analytics community, believe is clearly a substandard product with flaws that are most basic in nature and completely avoidable. It will be up to them to decide whether some harsh, but in my opinion valid, criticisms influence whether they seek my (or anyone’s) advice or not. What I do know is had the NHL sought out such advice previously I highly doubt I would have had to write this post. I really hope I can write a more positive review after the next phase is released in April.

Comments are closed.