DatsyukToZetterberg
Alligator!
Hi everyone. I am back again this season with my “DTZ Projections”. I apologize for the slight delay in releasing them; I wanted to see if Yahoo would make any more changes to player positions before finalizing the projections. However, it appears that there won't be any changes for the next little while.
If you don’t want to read about any of the changes to model and just want to see the projections, you can access the projections below:
While some players still have unconventional positions, like Kyrou or Verhaeghe being labeled as C's only, I chose not to rely on last year's positions. Last season showed that certain players might not regain their previous positions, as Hintz/Hertl never regained their LW eligibility. I'll make sure to update the sheets when Yahoo's updates positions.
A Brief Overview of the Model
The DTZ projections represent my own forecasts. I aimed to create projections that are fully adjustable. These sheets allow you to modify a player's GP or game state specific TOI, and the projections will adapt accordingly. Historically, my model has been conservative in terms of point totals, but does well in peripheral categories such as hits, blocks, and FOW, among others.
The goalie model is based entirely on the Marcel method, which employs a simple weighted average of a goalie's recent performances. I've adjusted the GP stat for certain goalies to account for who is the expected Starters/1As/1Bs/Backups. Although this sheet remains customizable to suit your preferences, I advise taking the goalie model with a pinch of salt.
Notable Changes This Season
I've refined the model to better incorporate the uptick in scoring since 2017, especially in recent years. A significant alteration is that the importance of expected statistics has diminished slightly compared to before. As mentioned earlier, I anticipate that this adjustment will result in slightly higher forecasts for scoring-related categories. In testing, this updated version has outperformed the previous iteration by a healthy margin. That being said, I expect this model to still be one of the more conservatism models.
Changes to the sheet
If you don’t want to read about any of the changes to model and just want to see the projections, you can access the projections below:
- Excel Version: You can download it from Dropbox. After the sheet loads, click on "Download" in the top left corner. Please note that enabling macros is necessary for the TOI/GP reset buttons to function. Here's a guide from MS to help with that.
- Google Sheets Version: To use this version, go to the top left corner and click on "File" -> "Make a Copy." Keep in mind that using the TOI/GP reset buttons might require granting certain permissions for the macro to operate.
While some players still have unconventional positions, like Kyrou or Verhaeghe being labeled as C's only, I chose not to rely on last year's positions. Last season showed that certain players might not regain their previous positions, as Hintz/Hertl never regained their LW eligibility. I'll make sure to update the sheets when Yahoo's updates positions.
A Brief Overview of the Model
The DTZ projections represent my own forecasts. I aimed to create projections that are fully adjustable. These sheets allow you to modify a player's GP or game state specific TOI, and the projections will adapt accordingly. Historically, my model has been conservative in terms of point totals, but does well in peripheral categories such as hits, blocks, and FOW, among others.
The goalie model is based entirely on the Marcel method, which employs a simple weighted average of a goalie's recent performances. I've adjusted the GP stat for certain goalies to account for who is the expected Starters/1As/1Bs/Backups. Although this sheet remains customizable to suit your preferences, I advise taking the goalie model with a pinch of salt.
Notable Changes This Season
I've refined the model to better incorporate the uptick in scoring since 2017, especially in recent years. A significant alteration is that the importance of expected statistics has diminished slightly compared to before. As mentioned earlier, I anticipate that this adjustment will result in slightly higher forecasts for scoring-related categories. In testing, this updated version has outperformed the previous iteration by a healthy margin. That being said, I expect this model to still be one of the more conservatism models.
Changes to the sheet
- The “Skater Projections” has an “Unadj VOR” column. This is included because of the current positions on Yahoo. It offers an alternative to the adjusted VOR which will likely be outdated after the 1st month of the season and Yahoo updates positions.
- The "Master Ranking" sheet features a "Pos Rank" (Positional Rank) column. This column showcases the rank of each player within their positional cohort. If a player holds eligibility for multiple positions, the sheet employs the position associated with their VOR (for example, JT Miller is treated as an RW).