I didn't realize you had quantified snow preservation
I have always taken snow preservation into account in making subjective recommendations. Note the last column in the second table on my regional pages "Best Time to Ski." http://18.104.22.168/~bestsnow/ncconet.htm This dates back to my original Powder Magazine cover story in October 1995. And everyone here should know this because of the that recur every time someone inquires about taking their March vacation in Jackson Hole.
The quantification comes from Chris Steiner's zeal to quantify all aspects of ski resort recommendations for his Forbes feature and the ZRankings website. Chris purchased my data in September 2013 and rolled out the website a couple of months later. The Snow Score in that first cut had a laundry list of flaws, and I insisted that if he was going to use any of my data, that he needed to give me control of the Snow Score algorithm. One of the flaws was that there was no provision for snow preservation, so I designed that a year ago. The redesigned Snow Score page has been on ZRankings since early last year, but was not used as an input to the overall PAF until the new Forbes feature was published last month. I do not know how Snow Score is used in the final PAF algorithm though.
It would be really cool to be able to select the week or month of the ski season and see the rankings for the particular time period. I'd think you could probably do that with your existing data by varying regression coefficients for elevation and exposure for each month/week of the season, in addition to taking into account when the snow is likely to fall at each area. I know you've already started down that path.... it would be great to see you expand on this so it dynamically scores all ski areas for a selected time period:
I don't think there's enough data to do that at the individual resort level on a widespread basis. You'll note that Colorado page is fairly generalized, with me estimating a handful of specific places as illustrative examples. To do it right you need input from local skiers or TR's over a decade or so. So I only had a spreadsheet in that format for Southern California, where I live and knew that tracking weekly conditions was the only way to reflect the frequent melt/freezes here. This is also necessary for the Northeast, and thus it took until 2010 or so to have a decade of online TR's to analyze in Vermont. Craig Morris' website http://www.redtree.com/far provided detail data from 1996-2008 at Fernie. My local friends in Utah have detail TR's since 2005. I had tracked Mammoth nearly as well as SoCal so the Sierra page was easy.
The ZRankings website has individual resort detail pages but they are a work in progress. I have begun writing customized snow comments for those. Here's Copper Mountain as an example, with Steiner's detail take on the area, followed by my comments on the snow: http://zrankings.com/resort/show?id=42-copper-mountain
Steiner has not done that many individual area detail write-ups. So here's Steamboat, where he has not yet done a write-up but I have: http://zrankings.com/resort/show?id=150-steamboat-ski-resort
I have supplied snow write-ups for the top 40 areas overall and for snow, but Steiner has not published all of those yet. The individual pages were created on a programmed not customized basis, and it takes time to do these write-ups. Steiner is much younger than I am and has a real world IT job, so it's not clear how much effort he's willing to invest to make those pages better. The annual Forbes piece is what gives him high visibility, so I'm sure that's his priority.
On the Snow Score page there is also a Snowfall Score 3 columns over. The difference between the two is based upon snow preservation. For major western areas preservation adds on average 5-6 points to Snow Score. So if Snow Score is similar to (Steamboat) or lower than (Jackson) Snowfall Score that's a sign that preservation is mediocre or worse, and that late season trips rate to see a lot of melt/freeze conditions. At the other extreme snow preservation adds 16.3 points to Copper's Snow Score and 21.5 points to A-Basin's.
For the East it's somewhat different because the big hit in snow preservation is due in large part to rain incidence. So you don't want to draw the conclusion that March is worse than other months in Vermont because a key risk factor applies to all months, not just the late season.