Well I don't think any of those sites list HDSF as a percentage of overall shots. He probably plugged everything into a spreadsheet to get the numbers.
Natural statrick has HDSF, it's just not one of the column's that are displayed by default.
I'm not sure what to make of the % of total shots metric since there are so many potential reasons for it.
For example, a team that is better at getting point shots through and on net would decrease it's overall % of HD shots since it would be increasing the number of low danger shots.
5v5, we're 8th in SCF/60, 15th in xGF/60 but 30th in HDCF/60.
High danger chances are anything that score more than 2 in the NST model, you get 2 for a shot from home plate, a +1 if it's a rebound, or rush chance, and a -1 if it's blocked. 2 is a medium danger chance (scoring chances are medium danger + High danger)/
So, maybe more of our high danger chances are getting blocked (turning them into medium danger chances) rather than missing the net?
I much prefer the xGF models which have far more variables they look at, they tend to do a better job than Scoring chance models when it comes to predicting goals, but it is none the less interesting that the team has such a big disparity between scoring chances and High danger chances.
If you look at the ranks of teams in SCF/60 and HDCF/60 (SCF Rank - HDCF Rank), we have the biggest discrepancy in the league at -22, with only one other team in the double digit negatives. Pretty odd.