PH Formulae

Dear all Sailwave PH users,

Maybe it's time for us to put all our ways, methods and formulae for
calculating PH on the table for Colin to see.

Personally PH is very high on my wish list and if at all feasible I
would like to adopt a common universal method if there is such a thing
(a way of ducking the firing line!). Perhaps this very lively discussion
can be carried forward in the interests of possibly developing a
'standard' universally acceptable PH system - or are the needs &
conditions unique to each club/regatta so far apart that we can't come
to agree on one accepted method, or even three or four along the lines
of IMS, PHRF etc? To me it seems as though each club has their own
backyard hashed system, some better than others; we do and it was
certainly not developed by a mathematician or statistician so I doubt
it's robustness.

This is what I do: PH is calculated using the formula c=e*r where c is
corrected time, r is the rating and e the elapsed time. The PH value (r)
is determined anew for each race and weighted 3/4ths of the previous
races' rating plus 1/4 of the current races' rating to dampen chance
wild oscillations. [Each new entry to the series is allocated the same
rating as the current lowest rated skipper]. I go to a spreadsheet to
calculate (competitors elapsed time/winners elapsed time) which gives a
new r value (range 0<r<1), the winners' new r value will be 1. Note I
use the winners time, not an nth competitors time and there are no
clamps or exclusions of the top or bottom percent. This new r value is
weighted 1/4 to 3/4 with the old r value and is used for the next race
(input into Sailwave edit+competitor default rating; I use custom rating
system in scoring properties and input the equation e*r in the c= box;
e=c/r & r=c/e) .

Perhaps the formula in a Sailwave custom PH facility might look
something like this

((Previous rating * 3) + (Competitors elapsed time/winners elapsed
time))/4

Comments!!

I have graphed the changes of r over two seasons and note all the common
failings of a PH system (1) it discriminates against very consistent
sailors who never seem to win a handicap series; (2) the value of r only
changes by ~3% (max 5%) and (3) the problems with establishing an
initial PH value. My observations and conclusions are well described and
are more eloquently put in an article titled "Handicapping algorithms
for measured performance HCing of keel boats" by Rod McCubbin passed
onto to me by Huw Pearce.

Yours,
Rob

Nice one, Rob! Encapsulates a lot of the difficulties with PH, especially
how to encourage the very consistent sailors (up to the handicapper to get
the initial PH 'correct'?). I'd be happy to use (and more importantly, sell
to the sailors) a formula along the lines that Rob suggests (even better if
the 2 weighting factors could be series variables, rather than built in).
Mike

···

-----Original Message-----
  From: Rob Walker [mailto:rswalker@mweb.co.za]
  Sent: 09 October 2004 07:47
  To: sailwave@yahoogroups.com
  Subject: [sailwave] PH Formulae

  Dear all Sailwave PH users,

  Maybe it's time for us to put all our ways, methods and formulae for
  calculating PH on the table for Colin to see.

  Personally PH is very high on my wish list and if at all feasible I
  would like to adopt a common universal method if there is such a thing
  (a way of ducking the firing line!). Perhaps this very lively discussion
  can be carried forward in the interests of possibly developing a
  'standard' universally acceptable PH system - or are the needs &
  conditions unique to each club/regatta so far apart that we can't come
  to agree on one accepted method, or even three or four along the lines
  of IMS, PHRF etc? To me it seems as though each club has their own
  backyard hashed system, some better than others; we do and it was
  certainly not developed by a mathematician or statistician so I doubt
  it's robustness.

  This is what I do: PH is calculated using the formula c=e*r where c is
  corrected time, r is the rating and e the elapsed time. The PH value (r)
  is determined anew for each race and weighted 3/4ths of the previous
  races' rating plus 1/4 of the current races' rating to dampen chance
  wild oscillations. [Each new entry to the series is allocated the same
  rating as the current lowest rated skipper]. I go to a spreadsheet to
  calculate (competitors elapsed time/winners elapsed time) which gives a
  new r value (range 0<r<1), the winners' new r value will be 1. Note I
  use the winners time, not an nth competitors time and there are no
  clamps or exclusions of the top or bottom percent. This new r value is
  weighted 1/4 to 3/4 with the old r value and is used for the next race
  (input into Sailwave edit+competitor default rating; I use custom rating
  system in scoring properties and input the equation e*r in the c= box;
  e=c/r & r=c/e) .

  Perhaps the formula in a Sailwave custom PH facility might look
  something like this

  ((Previous rating * 3) + (Competitors elapsed time/winners elapsed
  time))/4

  Comments!!

  I have graphed the changes of r over two seasons and note all the common
  failings of a PH system (1) it discriminates against very consistent
  sailors who never seem to win a handicap series; (2) the value of r only
  changes by ~3% (max 5%) and (3) the problems with establishing an
  initial PH value. My observations and conclusions are well described and
  are more eloquently put in an article titled "Handicapping algorithms
  for measured performance HCing of keel boats" by Rod McCubbin passed
  onto to me by Huw Pearce.

  Yours,
  Rob

  -!- http://www.fastmail.fm/ -!- http://www.spampal.org/ -!-
http://www.sailwave.com/ -!-

  Convert to daily digest of emails send blank email to:
  sailwave-digest@yahoogroups.com

        Yahoo! Groups Sponsor
              ADVERTISEMENT

----------------------------------------------------------------------------
--
  Yahoo! Groups Links

    a.. To visit your group on the web, go to:
    http://groups.yahoo.com/group/sailwave/

    b.. To unsubscribe from this group, send an email to:
    sailwave-unsubscribe@yahoogroups.com

    c.. Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.

[Non-text portions of this message have been removed]

Sail 100 has a good working formula to look at.

···

_____

From: Rob Walker [mailto:rswalker@mweb.co.za]
Sent: Saturday, 9 October 2004 4:47 PM
To: sailwave@yahoogroups.com
Subject: [sailwave] PH Formulae

Dear all Sailwave PH users,

Maybe it's time for us to put all our ways, methods and formulae for
calculating PH on the table for Colin to see.

Personally PH is very high on my wish list and if at all feasible I
would like to adopt a common universal method if there is such a thing
(a way of ducking the firing line!). Perhaps this very lively discussion
can be carried forward in the interests of possibly developing a
'standard' universally acceptable PH system - or are the needs &
conditions unique to each club/regatta so far apart that we can't come
to agree on one accepted method, or even three or four along the lines
of IMS, PHRF etc? To me it seems as though each club has their own
backyard hashed system, some better than others; we do and it was
certainly not developed by a mathematician or statistician so I doubt
it's robustness.

This is what I do: PH is calculated using the formula c=e*r where c is
corrected time, r is the rating and e the elapsed time. The PH value (r)
is determined anew for each race and weighted 3/4ths of the previous
races' rating plus 1/4 of the current races' rating to dampen chance
wild oscillations. [Each new entry to the series is allocated the same
rating as the current lowest rated skipper]. I go to a spreadsheet to
calculate (competitors elapsed time/winners elapsed time) which gives a
new r value (range 0<r<1), the winners' new r value will be 1. Note I
use the winners time, not an nth competitors time and there are no
clamps or exclusions of the top or bottom percent. This new r value is
weighted 1/4 to 3/4 with the old r value and is used for the next race
(input into Sailwave edit+competitor default rating; I use custom rating
system in scoring properties and input the equation e*r in the c= box;
e=c/r & r=c/e) .

Perhaps the formula in a Sailwave custom PH facility might look
something like this

((Previous rating * 3) + (Competitors elapsed time/winners elapsed
time))/4

Comments!!

I have graphed the changes of r over two seasons and note all the common
failings of a PH system (1) it discriminates against very consistent
sailors who never seem to win a handicap series; (2) the value of r only
changes by ~3% (max 5%) and (3) the problems with establishing an
initial PH value. My observations and conclusions are well described and
are more eloquently put in an article titled "Handicapping algorithms
for measured performance HCing of keel boats" by Rod McCubbin passed
onto to me by Huw Pearce.

Yours,
Rob

-!- http://www.fastmail.fm/ -!- http://www.spampal.org/ -!-
http://www.sailwave.com/ -!-

Convert to daily digest of emails send blank email to:
sailwave-digest@yahoogroups.com

Yahoo! Groups Sponsor

ADVERTISEMENT

<http://us.ard.yahoo.com/SIG=1297pmeui/M=295196.4901138.6071305.3001176/D=gr
oups/S=1705065792:HM/EXP=1097390820/A=2128215/R=0/SIG=10se96mf6/*http:/compa
nion.yahoo.com> click here

<http://us.adserver.yahoo.com/l?M=295196.4901138.6071305.3001176/D=groups/S=
:HM/A=2128215/rand=812426531>

  _____

Yahoo! Groups Links

* To visit your group on the web, go to:
http://groups.yahoo.com/group/sailwave/
  
* To unsubscribe from this group, send an email to:
sailwave-unsubscribe@yahoogroups.com
<mailto:sailwave-unsubscribe@yahoogroups.com?subject=Unsubscribe>
  
* Your use of Yahoo! Groups is subject to the Yahoo!
<http://docs.yahoo.com/info/terms/> Terms of Service.

[Non-text portions of this message have been removed]

Sounds good Rob so here goes :-))

Our club uses an Excel spreadsheet devised by the handicapping group for the Forth area. In it's simplest form it calculates the Standard Corrected Time for each race or fleet within the race and then calculates out each boats achieved performance.

Their is a bit of statistics to go along with how it does this and a better explanation can be found at

http://mysite.wanadoo-members.co.uk/fyca/Reference/FYCA_YR2Binstructions.pdf

You can also download a copy of the spreadsheet and have a look at it for yourselves.

From this we take the achieved performance subtract the original handicap and get a performance factor. We then divide this by four and carry these numbers forward for four races. After 4 races these numbers are averaged and the handicap is adjusted by this average either up or down to take out the swings that some races create.

The spreadsheet will throw out any realy fast or slow race results and we will not use these for our calculations.

As with all the systems it seems to work for us.

Trevor Mackay
Sailing Secretary
Cramond Boat Club

Rob Walker wrote:

Dear all Sailwave PH users,

Maybe it's time for us to put all our ways, methods and formulae for
calculating PH on the table for Colin to see.

<SNIP>