Given the way that the picks have gone lately, it only seemed appropriate to make some minor modifications to the system in an attempt to improve upon its performance. While it is true that all systems will probably run both hot and cold once in a while, it never hurts to revisit certain fundamentals to see if they can be improved upon. This particular system works by finding stats from past games of similar situations, and then using those stats to build a prediction, and so, in the latest set of updates, I’ve decided to make a few tweaks to what exactly it should be looking for.
With regards to travel, the system will now look at travel between different geographical areas. For example, a game in Vancouver followed by a game in Anaheim will now be labeled “NW_SW”, and the system will now factor in the stats from all previous instances of “NW_SW” for that specific team once it comes up again.
With regard to opponent type, the system will now use SOG stats instead of GF/GA stats to guage the play-style of the opposition, and the type of game which will likely be played.
Also, with regards to rating the quality of picks (from Gold, Silver, and Bronze, to the bottom of the barrel) there will now be more weighting placed on matchup edges, and less weighting on pure projected value. This will hopefully compensate for the possible margins of error which may exist in projected win percentages.
As for those projected win percentages (which are the teams’ estimated chances of winning a game), I think I will also be tweaking the formula for calculating them, so that the computer stops seeing value in almost every single dog on the board. This tweak will have to wait until tomorrow though.
Tuesday’s Sheet: (Goalies updated at 12PM PT)
Yeah, huge dogs always seem to be highlighted, so no surprise what the picks are today… Just to give an idea of what kind of implications a +150 dog has though, a team getting +150 only needs to win their game about 41% of the time to be profitable. There are only 5 teams in the league right now that have lower than .400 win percentages. Furthermore, +175 underdogs only need to win 37% of the time to be profitable. Even the Maple Leafs have a higher win% than that… I’ll still tweak the formula a bit during the day to make it a bit tighter though (less underdog-crazy).
Agree with the Gold pick? Disagree with it? See a betting line on the board that seems wrong? You know what to do (fill up the comment section).