rulefan
Tour Winner
They have.
Fundamentally it is now about expected field scores vs actual. Previously it was about actual scores only.
PCC review
- Thread starter Junior
- Start date
Fundamentally it is now about expected field scores vs actual. Previously it was about actual scores only.
RRidges
Active member
I'd bet you can't explain the difference because, to me, there is none - save wording! It's always been CSS that's used, so actual scores only.
And btw. How does CR+PH (expected) vs Day's Results differ from SSS vs Actual Values (from which CSS was calculated)?
wjemather
Well-known member
CSS was based on a target score (playing to handicap, per SSS after NDB adjustments) and buffer zones that has no basis in or correlation with actual scoring patterns; PCC is based on expected scoring ranges that are based on actual scoring patterns.
.
RRidges
Active member
That clarifies things somewhat - thanks. But I'd still expect the new calc to recognise that virtually all scores are at the high (poor) end of that range, then determine the adjustment required for it to become closest to the middle(?) of the expected value. It doesn't seem to be doing the 2nd bit particularly well. Or was the CSS calc too sensitive, so not a great stat to compare?
wjemather
Well-known member
PCC is triggered based on the ratio of the proportion of players in, above and below their expected range, and provides an adjustment for when scoring is significantly better/worse than expected; it isn't really designed to normalise scores.
There are several reasons for the higher sensitivity of CSS, including: for low handicaps, buffer zones were very small so players could return half decent/average scores and be outside their buffer; and for higher handicaps, due to there being no accounting for Slope, at most playing handicaps were lower than they should have been and so returned nett/Stableford scores weren't as good as they perhaps should have been, resulting in an increased number of players outside their buffer.
Last edited:
D-S
Well-known member
Didn’t the ‘smaller’ buffer zones for lower handicap golfers correlate with the narrower expected scoring range for lower handicappers and the wider buffer zones for higher handicappers did the same? It is just a question of where you set the ‘expected scoring limits’.
I agree that higher handicappers typically now have more shots than before due to slope but this of course is a contentious enough subject for other discussions as well.
I agree that higher handicappers typically now have more shots than before due to slope but this of course is a contentious enough subject for other discussions as well.
wjemather
Well-known member
Not really, or at least not anything like closely, since only a relatively small percentage of scores are that narrowly distributed for even the most consistent of (former) cat1 players.
Alan Clifford
Active member
One would expect the algorithm to be hacked eventually. Except no-one could be arsed to make the effort because on a scale of 1 to 10 of important things in the world to be hacked, this rates at about minus 5.
D-S
Well-known member
So what is the likely scoring range now for let’s say 0-5 HI players now according to the PCC algorithm?
RRidges
Active member
Indeed CSS was more sensitive than PCC - which is why folk are complaining, though I don't really consider myself one of them - just a bit puzzled, but your statement re 'range of scores' explains most of the difference.
I'd disagree with you re the bold bit, as for stable players the playing/actual handicap under Congu should have been pretty much the same as Playing Handicap under WHS. It was merely the starting HI that is lower under WHS. That was certainly the case for me coming from Aus system to Congu and pretty similar experience with Americans, either in UK or in USA until I got used to that level.
I presume the calc change will be worldwide.
There's a thread on GolfWRX forum that documents the distinct reduction of PCC adjustments in US and Canada since WHS was implemented, so it'll be interesting to see what effect the adjust calc makes. It's actually largely irrelevant to me, as I'm a fair weather golfer anyway, though playing in temps in the mid 90sF does affect energy levels and post round thirst!
Last edited:
wjemather
Well-known member
The expected ranges are one of the unknowns of the PCC, although it is known that they are calculated individually for each index.
However, based on analysis of some actual handicap records (obviously an insignificant fraction of number used by the WHS development team) and taking at a reasonably sized median percentile (i.e. at least 60-70%) of scores, I'd estimate the expected range to be roughly from playing to handicap to 4 (or more) over for them, and 6 (or more) over for higher handicappers.
It's only a rough estimate but gives a good idea of how much worse the scoring needs to be to trigger a positive PCC than it did previously with CSS.
wjemather
Well-known member
For our course (Slope 127 & 126), we generally saw parity in playing/course handicaps between UHS and WHS around the 8-14 handicap range, with higher handicappers getting more strokes under WHS, and lower handicappers getting fewer.
rulefan
Tour Winner
Thanks, it is something I didn't know before. Is that the rounded or unrounded index?
Incidentally, just re-reading 5.6 it struck me that each set of tees has its own PCC and presumably for each gender.
wjemather
Well-known member
In 5.6/1 it is worded as being calculated for each eligible player, so by extension that means each HI and each set of tees being played.
RRidges
Active member
That seems close enough to to my 'pretty much', though maybe that range is a little lower than I'd have expected. The major difference between the systems in this area is not the HI vs 'hard' handicap, but the average of best 8 from last 20 method of calculating the HI in the first place, which is golfer dependent - streaky players end up lower than consistent ones under WHS. When UHS was in place, we normally added 2-3 shots to the HIs of US guys in that range, or a bit more, that came over (I work for a US consulting company, so there were quite a few) to equate to UK UHS handicaps unless the actual Slope and CR values were known. That normally worked out pretty well for the standard of courses we played.
But alas, we've diverted from PCC v CSS area to the wider WHS v UHS area.
I'll await the reaction to the change to PCC calc from actual results. It's the considerable difference in the purpose of the calc that seems to be the cause - and apparently prompted by Aus numbers/calculations. Don't blame me guv!
Last edited:
badgergm
Newbie
And what would be the problem with this?
RRidges
Active member
That reply is a bit ambiguous. Are you querying the 'competing package' or the 'claiming it is exactly the same'?
rulefan
Tour Winner
badgergm
Newbie
Maybe I dont understand the model here. Who owns the specification of the algorithm? I’m sssuming that it is the handicapping body. And that they pay a software company to implement that algorithm on a system which runs overnight calculating all the PCCs for clubs based on the scores submitted the previous day.
The IPR would rest with the client, not the software company. Who else is going to buy the software?
wjemather
Well-known member
Being able to recreate PCC is pointless without having access to a substantial amount of daily scores in order to make the calculations meaningful. It's unlikely that would ever be the case for a third party unofficial handicap provider.
