Want create site? Find Free WordPress Themes and plugins.

Market Inefficiencies and the NHL

Background

When we at Money Puck Systems first started valuing NHL players in about 2015 we started by developing a robust tool for the job. Based on our experience in finance we knew that the first step to developing an intrinsic value for players would be to do a good job forecasting the key inputs (so we built this awesome tool to help us do that!). We assumed that like in finance you would do your best to estimate these future things like the underlying performance of the asset (in finance that refers typically to a business and in hockey it’s a player), inflation rates (or salary cap changes for hockey), etc. and despite the fact that you’d never be right in your forecast the exercise would get you reasonably close and certainly closer than not trying at all…

How It Really Works

What got us started in our venture was that our new model ‘felt’ pretty close just by plugging in some historical data on player performance and when there seemed to be differences from actual compensation and what our model said it also ‘felt’ like what our model said was better than what was observed (imagine what our model said versus the Phaneuf contract at the time for example). The reason it worked so well with just historical data though was because that is how the actual market is setting the prices. We don’t feel it’s the correct approach and that you’d be better served by with a solid forward looking forecast but if you want to accurately predict signings come July 1st don’t start with a forecast – just look at how the player did last year. In fact, when we release our estimated contract signings in a month or so that’s exactly what we’ll do. We’ll use our model a little bit but more often we’ll be leaning on a formula we built using a basic linear regression analysis with the only factor being last season’s points.

If you’re curious how well that works here’s how well it fits the data for select UFA wingers signed in the last 3 years:

Pic1

To get a little closer you can obviously tweak for special circumstances (i.e. that dot just above 6% and 0.4 points/game is Tom Wilson) but on the whole we have been very successful in predicting contracts just by creating 6 charts like this (all the combinations of UFA, RFA, Centers, Wingers and Defencemen). To demonstrate our success our 2018 predictions looked like this:

Pic2

What’s the Problem?

Please excuse that so far this just feels like a Money Puck Systems ad but the point is that while we can pretty well predict what players will be signed for we can also anticipate how good or bad these contracts will look over time. It’s really as simple as estimating the player’s future performance versus what the market price of that player will be. A really good example is the UFA signings of 2016. All the big contracts that year were right in line with what should have been expected and despite what we know now I wouldn’t call any of them ‘overpays’ – they were roughly the market price for those players at the time. So which ones were those again:

– Kyle Okposo ($6 million/year for 7 years)
– Milan Lucic ($6 million/year for 7 years)
– David Backes ($6 million/year for 5 years)
– Loui Eriksson ($6 million/year for 6 years)
– Andrew Ladd ($5.5 million/year for 7 years)
– Frans Nielsen ($5.25 million/year for 6 years)

Anybody interested in acquiring those contracts now?

Of course no one wants these contracts anymore (except maybe Nielsen) and the reason is obvious: all of these players experienced declining production in the 3 years since these contracts were signed. Now to be clear this isn’t just hindsight bias and I’m not just selecting contracts that make my case. It’s been shown time and again that as players age their play eventually begins to deteriorate. We can debate what age that is and which metrics to use for performance but it should be well known that most of the time that you sign a player in their late 20s or early 30s to a long term contract – those players will not continue to perform at the rate they currently do. When it comes to paying them though, NHL teams effectively pretend that aging curves don’t exist or that their guy will be different. If that weren’t true you wouldn’t see such a high correlation between last season’s points and what a player is paid. You’d also see us needing to use our forecast model in order to get these contract estimates right – we don’t. In fact as far as I know none of us well known for predicting signings use any future looking inputs to do it…

How to Benefit?

What would a smart team do with this information and how can it be used to your advantage? Well first do not sign players in their late 20s or early 30s to long term, high value contracts. You actually have to forecast what their performance looks like for every year of that potential contract and apply an aging curve to their performance. Almost always that will mean some other team will pay more than you and you don’t get that shiny new UFA, but unless there’s a very good strategic reason to do it (like guaranteeing cup or two in the first part of the contract) you are simply going to handcuff your team in future years and the GMs who can resist that temptation will be much better over the long term.

Second, use the aging curve to your advantage with young players. Just like their older counterparts the younger players are also paid based on what they did last season. The difference is that they are likely to get better over time and become a bargain as the contract ages. An exaggerated example is the Jonas Brodin contract but the idea is to lock these players in for as long as possible before they hit their peak. Of course, this is always a gamble and some players won’t work out, but I’d certainly rather take this gamble than betting on that my veteran players won’t age at all over 6 or 7 years.

Combining both of those ideas, it is generally advisable to trade your late 20s and early 30s veterans for younger players. For whatever reason current GMs simply overvalue these players and you should be able to get maximum trade value for them even though they are past their peak and typically about to fall off a performance cliff (imagine what Jonathan Toews’ trade value was in 2015 just before he went on to put up 3 consecutive seasons with less than 60 points – although I must admit he surprised with an 81 point season last year). This requires great pro scouting to make sure you aren’t constantly trading for ‘magic beans’ and undoubtedly it will be deeply troubling for fans who have grown to love these players but it’s the smart play.

Next, teams should ‘over-ripen’ their prospects. Detroit was famous for it during their 25 year playoff streak and Ken Holland has been quoted as a proponent of the ‘draft-and-over-ripen’ philosophy. Now I’m not an expert on player development and what works best for players. If someone can put together good research to suggest that too much time in the minors is detrimental to a players’ development I am open to changing my view but given Detroit’s success during this era I do not believe an extra year or two in the AHL significantly harms a player. What it does do from a cap perspective though is it moves the players’ cheapest years, their entry-level years, to a more productive part of their performance curve. To see it visually consider Hawerchuk’s points-per-game chart and imagine getting 3 years of your player at $1 million/year from 23 to 26 instead of from 21 to 24. Clearly there’s a lot more value and the player would command a lot more money from 23 to 26 but by over-ripening your players the team can benefit from this aging curve instead of the player.

pic3

Putting It All Together

Lastly, build for windows. In order to be an elite team the truth is that the value of your roster needs to be significantly more than what they are actually paid and typically quite a bit higher than what the salary cap would allow. As an example, we estimated what the 2016-17 Pittsburgh Penguins team would have been worth on the free agent market and despite a $70.4 million salary cap we estimate that team to be worth closer to $100 million! This is an incredible difference and would probably still be hard to achieve even if you did everything described above. It can however be achieved if you create a ‘window’ by having your biggest contracts all end at about the same time.

The reason for this is that the salary cap is generally increasing over time but salary cap hits stay flat. What this means for your best players on the biggest and longest deals is that even if those contracts are ‘fair’ (i.e. their market value over the years were to equal their contract total) those contracts will actually be ‘overpays’ early and ‘underpays’ at the end. So the idea here is to build a young core, sign them just before they peak to long term deals and then plan to add a bunch of other value to your team for the back half of those contracts (even if it sometimes means breaking the other rules on short term deals because you probably only have about 4 years to win that cup before you’ll need to start trading that aging core).

Let us know your thoughts!

This article was written using the MoneyPuck calculator. Try it out for yourself!
MoneyPuck Login: MP_User
Password: nhl123

Did you find apk for android? You can find new Free Android Games and apps.
Market Inefficiencies and the NHL

Leave a Reply

Your email address will not be published. Required fields are marked *