We sweated it out a bit last night with the Florida Panthers, but they pulled through for us and capped off another vastly profitable night in the NHL betting streets. Netting 12.35 units just adds to our very successful season so far - one so good it would be easy to just coast through the next few months with the same practices.
I'm very analytical, as you can probably tell, and much like the underlying numbers I reference in my hockey analysis I also have underlying numbers on our betting success. These underlyings, while good, showed room for improvement in some areas. Since I'm never complacent, this called for some adjustments to the model.
You could probably tell based on yesterday's numbers, but different measures were taken to properly pinpoint a projected goals number for each team playing last night. This required slight rate adjustments based on more intricate details, and resulted in a healthy balance of suggested overs and unders, as well as -1.5 and +1.5 pucklines.
No longer is there simply an "opponent adjustment", rather, an opponent adjustment accounting for projected line matchups. The home team in the NHL gets the option for last change after a stoppage, allowing them to put the players they want on the ice against an opponent's line. Since some lines are better defensively than others within a team, this more intricate detail should give us a truer projected goals number for each player.
Similarly, certain players on a team are better offensively, and can lift a new linemate's production. This was always accounted for this season in the model, but now has a truer correlation percentage to each teammate. Not only should improving correlation percentages improve accuracy of point outputs, but also allows us to create a better correlation tool than the one on the site last season. Look for that tool to be released this week.
If you've followed PuckLuck since its inception, you know goalie impacts have been significant within the model. That should still prove true, due to the importance of the position, but the nature of that impact should also be more exact. Goaltenders perform differently in all situations, and their actual goals saved above expected numbers will now be fluid with who is on the ice for both teams. That is now accounted for dynamically within the model.
You can scour the internet for other NHL projection models, and I'm confident you won't find anything in the public-sphere that is this in-depth. Most game projections take generic team numbers and don't account for players in and out the lineup and in some cases even starting goalies. No fantasy projections out there consider the exact players that will be on the ice for a certain amount of time, including line matchups, starting goalies for each team, correlated teammates, and projected powerplay and penalty kill time based on penalty drawers and takers.
These things all have an undeniable impact on the game, and accounting for them sets us apart from everyone out there - including the sportsbooks. We had a very successful first night using the new dynamics, and I will continue to monitor throughout the season to make sure our returns are top-notch.
With that, let's see what edges the model has produced for tonight: