Since I was the culprit for this formula let me add a couple of comments
- It was *derived* through trial/error to fit all available data last year, as such may not be accurate down to 1-2 points which is a 0.2% error margin.
- The original formula was 5VR + DM + QR + 1.5AR. Expressed in more exact % terms would be 58.823%VR + 11.765%DM + 11.765%QR + 17.647%AR so I rounded it to 58%VR + 12%DM + 12%QR + 18%AR.
lowland7's discrepancy could be due to this rounding.
Regarding WSU's big jump in non-rural cutoffs, with the 90%ile score being ~100 marks higher = 25 marks/section it's not a big surprise the formula's cutoff has risen ~25 marks.