I didn’t really plan to discuss predictability in this series about talent distribution, since the two are not directly connected. However, I was asked about it, and it does open up for a good discussion about descriptive vs. predictive modelling. I recommend reading part 1 before you read this piece. Descriptive vs. Predictive models TheContinue reading “Talent Distribution – Predictability (Part II)”
Tag Archives: p-sGAA
Season Projections 20/21 – Updated
The previous season projections can be found here. There are a few important things to be aware of when interpreting the model. Important notes and updates: The previous projection had a mistake in the goaltender projections, but this is fixed now. The updated model is based on lineups (including taxi squad) from http://www.capfriendly.com/depth-charts. Players onContinue reading “Season Projections 20/21 – Updated”
Season Projections – First draft
Here’s my first season projections. The plan is to do an analysis of each division in greater detail, but there are still a few changes I would like to make. I will get back to that later. Updated data: Before we get started, its important to note that I have updated the data behind theContinue reading “Season Projections – First draft”
Building the game projection model
Abstract: In the previous article I laid the groundwork for my game projection model. Now I will use the actual starting lineups and add in season adjustments using the ELO ranking principle. The data used in this article is from the 2016/2017 season to the 2019/2020 season. Finding the starting lineups: I used Evolving-Hockey’s ShiftContinue reading “Building the game projection model”
Converting season projections to game Projections
Abstract: The goal in this article is to convert my projection model (Read more here and here) from season projections into game projections. From season projections to game projections: The model as it stands right now uses p-sGAA to project team strength. I want to convert that into Win%, which is the projected number ofContinue reading “Converting season projections to game Projections”
Projecting future results – Goaltender edition
Abstract: In the last article, I tried to project future performances of the NHL skaters. In this piece I will add goaltender projections, so we can see how the first draft of the projection model performs. What if all goaltenders were average? Before we get to the goaltender projections, let’s try to assume every goaltenderContinue reading “Projecting future results – Goaltender edition”
Projecting future results – Skater edition
Abstract: I have previously introduced the sGAA model (See here), which is primarily a descriptive model. Now I will introduce a predictive counterpart – projected sGAA or p-sGAA. It’s based on 3 years of weighted data. sGAA describes past performances while p-sGAA projects future performances. The projection model: The model is basically based on 3Continue reading “Projecting future results – Skater edition”