Problem with Corrected Time Precision

mtownsend · 26 June 2010 14:09

PROBLEM 1 Corrected Time Precision

The corrected times of Rebel Yell and Grand Illusion are not being calculated sufficient precision to resolve the corrected time tie between the boats. In the 2005-2008 Racing Rules of Sailing rule A3 was changed to remove to the phrase corrected to the nearest second so that the individual rating rules now determine the precision of the corrected results. By rounding the result to the nearest second a tie is being created that should not exists. There is no method in SailWave to specify the precision to which you want the corrected time calculated.

SAILWAVE CALCULATION

Rank

Boat

Nat

SailNo

HelmName

Rating

Start

Finish

Elapsed

Corrected

BCE

1

Wasabi

USA

55544

  Dale Williams

-21

14:50:00

16:07:45

1:17:45

1:23:40

0:00:00

2

  Rebel Yell

USA

52005

  David O. Team

-84

14:50:00

16:00:02

1:10:02

1:24:09

0:00:24

2

  Grand Illusion

USA

97

  Edward McDowell

-69

14:50:00

16:01:59

1:11:59

1:24:09

0:00:25

4

Staghound

USA

69152

  Alec Oberschmidt

-57

14:50:00

16:03:39

1:13:39

1:24:17

0:00:32

5

  Its OK

USA

7249

  Tres Gordo

-48

14:50:00

16:04:59

1:14:59

1:24:28

0:00:43

5

Locomotion

USA

46860

  Ed Feo

-21

14:50:00

16:08:02

SCP

1:23:59

0:00:17

7

Flash

USA

62152

  Mick Shlens/Mark Jones Shlens/Jones

-87

14:50:00

16:01:02

1:11:02

1:25:49

0:01:47

8

Cazador

USA

28385

  Ernie Pennell

-78

14:50:00

16:02:28

1:12:28

1:26:07

0:02:03

9

Peligroso

MEX

55555

  Lorenzo Berho

-99

14:50:00

15:59:50

1:09:50

1:26:18

0:02:07

10

  Emirage II

USA

48003

  David Bassett-Parkins

-30

14:50:00

16:11:04

1:21:04

1:28:34

0:04:29

11

Limit

AUS

98888

  Alan Brierty

-138

14:50:00

15:56:30

1:06:30

1:28:45

0:03:49

EXCEL CALCULATION

Boat

Rat

  TCF A=650 B=625

Start

Finish

Elapsed

Corrected

1

Wasabi

-21

1.07616

14:50:00

16:07:45

1:17:45

1:23:40

1:23:40.281

2

  Rebel Yell

-84

1.20148

14:50:00

16:00:02

1:10:02

1:24:09

1:24:08.614

2

  Grand Illusion

-69

1.16906

14:50:00

16:01:59

1:11:59

1:24:09

1:24:09.191

4

Staghound

-57

1.14437

14:50:00

16:03:39

1:13:39

1:24:17

1:24:16.954

5

  Its OK

-48

1.12652

14:50:00

16:04:59

1:14:59

1:24:28

1:24:28.198

5

Locomotion

-21

1.07616

14:50:00

16:08:02

SCP

1:18:02

1:23:59

1:23:58.576

7

Flash

-87

1.20818

14:50:00

16:01:02

1:11:02

1:25:49

1:25:49.257

8

Cazador

-78

1.1883

14:50:00

16:02:28

1:12:28

1:26:07

1:26:06.728

9

Peligroso

-99

1.23574

14:50:00

15:59:50

1:09:50

1:26:18

1:26:17.757

10

  Emirage II

-30

1.09244

14:50:00

16:11:04

1:21:04

1:28:34

1:28:33.613

11

Limit

-138

1.3347

14:50:00

15:56:30

1:06:30

1:28:45

1:28:45.462

RRS 2009-2012

A3 STARTING TIMES AND FINISHING PLACES

The time of a boats starting signal shall be her starting time, and the order in which boats finish a race shall determine their finishing places. However, when a handicap or rating system is used a boats corrected time shall determine her finishing place.

RRS 2001-2004

A3 STARTING TIMES AND FINISHING PLACES

The time of a boats starting signal shall be her starting time, and the order in which boats finish a race shall determine their finishing places. However, when a handicap system is used a boats elapsed time, corrected to the nearest second, shall determine her finishing place.

Mark Townsend
s_mark_townsend@hotmail.com

Art_Engel2 · 26 June 2010 23:34

I think you might be mixing two issues - whether SailWave can be set to round to a precision defined by the user and what precision SHOULD be used. I think I might take issue with your statement "By rounding the result to the nearest second a tie is being created that should not exists [sic]."

I would agree that it might be helpful if SailWave could be set to round to say tenths or hundredths of a second. However, I think that would only rarely, if ever, be appropriate. Hence, I wouldn't put it high on a priority list of features for a scoring program (or even include it on such a list).

The RRS generally don't specify stuff that can be resolved without a special definition. For instance, they don't specify what it means to "tack" so the commonly understood definition applies and does an adequate job. [Being "on a tack" is a little different because of sailing "by the lee" so a special definition IS required for that.]

In the case of rounding corrected times, the RRS used to specify that all corrected times should be rounded to the nearest second - even if the elapsed times were taken to a tenth of a second (which is what you get when times are taken with a stopwatch that displays so-called decimal hours [9.545 hours]). But, rating/handicap systems can specify the precision of corrected times plus we have the standard scientific and engineering conventions so the "rounding" language was deleted as unnecessary.

As a scorer, with no instructions from the RRS I would decide how to proceed in the following order:

1. Do what the rating/handicap system says
2. Use scientific and engineering conventions
3. Do what most other scorers are doing (i.e., follow tradition)
4. Do what your scoring program allows

So, if rating or handicap system is silent then I would use the standard conventions. The standard convention is to round final calculations to the precision of the least significant component. So, if somewhere in your calculation formula you are using whole seconds (but tenths or hundredths for other components) then final numbers should be rounded to the nearest second. Finish times are typically taken to the nearest whole second so that would dictate corrected times rounded to the nearest whole second (with standard PHRF-type calculations rating are given in whole seconds as well so that is a 2nd place where the minimal precision is nearest whole seconds).

If virtually everyone is using some rounding convention then I would consider that as a possible alternative to the standard conventions. Since most scoring is rounded to the nearest whole second that would be a 2nd reason to round to the nearest whole second.

Finally, if your scoring program only allows a certain kind of rounding then I would simply use that. So long as all boats are scored the same way there isn't any unfairness and there shouldn't be grounds for any redress by competitors. From this perspective no method of rounding could ever be "wrong" but rounding to the nearest whole second would be BEST when finish times are recorded to the nearest whole second.

BOTTOM LINE: I think boats whose corrected time is the same when rounded to the nearest whole second SHOULD be tied (based on whole second finish times). However, it wouldn't break the RRS if such boats were scored as NOT tied based on a difference in tenths of a second so long as the corrected times of ALL boats were calculated to the same precision (all boats in the same class for the whole regatta or series, not all boats in other classes as well). In that regard, I think SailWave is just fine rounding corrected times to the nearest second since I would guess that 99.999% of finish times are entered to the nearest second. That is consistent with standard convention, usual scoring practice and tradition. I do NOT see a problem here.

Art Engel

S. Mark Townsend wrote:

PROBLEM 1 - Corrected Time Precision

The corrected times of Rebel Yell and Grand Illusion are not being

calculated sufficient precision to resolve the corrected time tie
between the boats. In the 2005-2008 Racing Rules of Sailing rule A3 was
changed to remove to the phrase "corrected to the nearest second" so
that the individual rating rules now determine the precision of the
corrected results. By rounding the result to the nearest second a tie is
being created that should not exists. There is no method in SailWave to
specify the precision to which you want the corrected time calculated.

RRS 2009-2012

A3 STARTING TIMES AND FINISHING PLACES

The time of a boat's starting signal shall be her starting time, and

the order in which boats finish a race shall determine their finishing
places. However, when a handicap or rating system is used a boat's
corrected time shall determine her finishing place.

···

Mark Townsend
s_mark_townsend@hotmail.com

Philippe_DE_TROY · 27 June 2010 07:10

The precision to the second was removed in RRS 2009-2012 to allow different precision especially for developping electronic (automatic) timing. However I think this is not a good choice. Automatic timing is based on on-board GPS sensor using fixed beacons, hence very good precision. But the sensor is on the cebtreline of the boat and using this for timing implies larger changes in the RRS like definition of boat starting and finishing a race...

Phil

···

----- Original Message ----- From: "Art Engel" <artengel123@earthlink.net>
To: <sailwave@yahoogroups.com>
Sent: Sunday, June 27, 2010 1:34 AM
Subject: Re: [sailwave] Problem with Corrected Time Precision

I think you might be mixing two issues - whether SailWave can be set to
round to a precision defined by the user and what precision SHOULD be
used. I think I might take issue with your statement "By rounding the
result to the nearest second a tie is being created that should not
exists [sic]."

I would agree that it might be helpful if SailWave could be set to round
to say tenths or hundredths of a second. However, I think that would
only rarely, if ever, be appropriate. Hence, I wouldn't put it high on a
priority list of features for a scoring program (or even include it on
such a list).

The RRS generally don't specify stuff that can be resolved without a
special definition. For instance, they don't specify what it means to
"tack" so the commonly understood definition applies and does an
adequate job. [Being "on a tack" is a little different because of
sailing "by the lee" so a special definition IS required for that.]

In the case of rounding corrected times, the RRS used to specify that
all corrected times should be rounded to the nearest second - even if
the elapsed times were taken to a tenth of a second (which is what you
get when times are taken with a stopwatch that displays so-called
decimal hours [9.545 hours]). But, rating/handicap systems can specify
the precision of corrected times plus we have the standard scientific
and engineering conventions so the "rounding" language was deleted as
unnecessary.

As a scorer, with no instructions from the RRS I would decide how to
proceed in the following order:

1. Do what the rating/handicap system says
2. Use scientific and engineering conventions
3. Do what most other scorers are doing (i.e., follow tradition)
4. Do what your scoring program allows

So, if rating or handicap system is silent then I would use the standard
conventions. The standard convention is to round final calculations to
the precision of the least significant component. So, if somewhere in
your calculation formula you are using whole seconds (but tenths or
hundredths for other components) then final numbers should be rounded to
the nearest second. Finish times are typically taken to the nearest
whole second so that would dictate corrected times rounded to the
nearest whole second (with standard PHRF-type calculations rating are
given in whole seconds as well so that is a 2nd place where the minimal
precision is nearest whole seconds).

If virtually everyone is using some rounding convention then I would
consider that as a possible alternative to the standard conventions.
Since most scoring is rounded to the nearest whole second that would be
a 2nd reason to round to the nearest whole second.

Finally, if your scoring program only allows a certain kind of rounding
then I would simply use that. So long as all boats are scored the same
way there isn't any unfairness and there shouldn't be grounds for any
redress by competitors. From this perspective no method of rounding
could ever be "wrong" but rounding to the nearest whole second would be
BEST when finish times are recorded to the nearest whole second.

BOTTOM LINE: I think boats whose corrected time is the same when rounded
to the nearest whole second SHOULD be tied (based on whole second finish
times). However, it wouldn't break the RRS if such boats were scored as
NOT tied based on a difference in tenths of a second so long as the
corrected times of ALL boats were calculated to the same precision (all
boats in the same class for the whole regatta or series, not all boats
in other classes as well). In that regard, I think SailWave is just fine
rounding corrected times to the nearest second since I would guess that
99.999% of finish times are entered to the nearest second. That is
consistent with standard convention, usual scoring practice and
tradition. I do NOT see a problem here.

Art Engel

S. Mark Townsend wrote:

PROBLEM 1 - Corrected Time Precision

The corrected times of Rebel Yell and Grand Illusion are not being

calculated sufficient precision to resolve the corrected time tie
between the boats. In the 2005-2008 Racing Rules of Sailing rule A3 was
changed to remove to the phrase "corrected to the nearest second" so
that the individual rating rules now determine the precision of the
corrected results. By rounding the result to the nearest second a tie is
being created that should not exists. There is no method in SailWave to
specify the precision to which you want the corrected time calculated.

RRS 2009-2012

A3 STARTING TIMES AND FINISHING PLACES

The time of a boat's starting signal shall be her starting time, and

the order in which boats finish a race shall determine their finishing
places. However, when a handicap or rating system is used a boat's
corrected time shall determine her finishing place.

Mark Townsend
s_mark_townsend@hotmail.com

------------------------------------

-!- http://www.yachtsandyachting.com/ -!- http://www.sailing.org/ -!- http://www.sailwave.com/ -!- On-Line Sailwave help...http://sailwave.com/help/HTML ~ Mark Townsend's Sailwave User Guide is available from http://www.abyc.org/upload/Sailwave_ABYC_User_Guide.pdf ~ Convert to daily digest of emails send blank email to sailwave-digest@yahoogroups.com ~ To unsubscribe from the SUG please send blank email to sailwave-unsubscribe@yahoogroups.com Yahoo! Groups Links

Art_Engel2 · 27 June 2010 22:53

The change was approved by the ISAF Racing Rules Comm. (and ISAF Council) in 2002 to become effective with the 2005-08 rulebook. The principal reason given was that it was more appropriate for such matters to be determined by the rating/handicap authority or the scorer, rather than the racing rules. "Automatic" timing was not mentioned and presumably was not a factor (or a only very minor factor).

There are plenty of obstacles to "automatic" timing. In recent years (after the change in 2002) ISAF has apparently started to explore the idea but so far as I know it hasn't gotten beyond the "thinking" stage. Others may have tried to make devices but I haven't heard of any success.

The "correct to the nearest second" concept was introduced into the rulebook in the 2001-04 edition as part of an extensive redraft of Appendix A that was done by one individual and adopted by ISAF (i.e., there was no formal submission). The idea of "corrected times" hadn't previously been mentioned by the racing rules at all. Within a year it was determined that the mention of "corrected times" was a good addition to the racing rules but the idea of legislating precision (i.e., how to round corrected times) probably was not. Hence, the "nearest second" concept was only in the racing rules for 4 years (and for 2+ of those years it was scheduled to be removed at the first opportunity).

Art Engel

Caveat - The "correct to nearest second" language was deleted as the result of a submission from the USA. I was a member of the US Racing Rules Committee at the time (and still am).

Philippe DE TROY wrote:

···

The precision to the second was removed in RRS 2009-2012 to allow different precision especially for developping electronic (automatic) timing. However I think this is not a good choice. Automatic timing is based on on-board GPS sensor using fixed beacons, hence very good precision. But the sensor is on the cebtreline of the boat and using this for timing implies larger changes in the RRS like definition of boat starting and finishing a race...

Phil

----- Original Message ----- From: "Art Engel" <artengel123@earthlink.net>
To: <sailwave@yahoogroups.com>
Sent: Sunday, June 27, 2010 1:34 AM
Subject: Re: [sailwave] Problem with Corrected Time Precision

I think you might be mixing two issues - whether SailWave can be set to
round to a precision defined by the user and what precision SHOULD be
used. I think I might take issue with your statement "By rounding the
result to the nearest second a tie is being created that should not
exists [sic]."

I would agree that it might be helpful if SailWave could be set to round
to say tenths or hundredths of a second. However, I think that would
only rarely, if ever, be appropriate. Hence, I wouldn't put it high on a
priority list of features for a scoring program (or even include it on
such a list).

The RRS generally don't specify stuff that can be resolved without a
special definition. For instance, they don't specify what it means to
"tack" so the commonly understood definition applies and does an
adequate job. [Being "on a tack" is a little different because of
sailing "by the lee" so a special definition IS required for that.]

In the case of rounding corrected times, the RRS used to specify that
all corrected times should be rounded to the nearest second - even if
the elapsed times were taken to a tenth of a second (which is what you
get when times are taken with a stopwatch that displays so-called
decimal hours [9.545 hours]). But, rating/handicap systems can specify
the precision of corrected times plus we have the standard scientific
and engineering conventions so the "rounding" language was deleted as
unnecessary.

As a scorer, with no instructions from the RRS I would decide how to
proceed in the following order:

1. Do what the rating/handicap system says
2. Use scientific and engineering conventions
3. Do what most other scorers are doing (i.e., follow tradition)
4. Do what your scoring program allows

So, if rating or handicap system is silent then I would use the standard
conventions. The standard convention is to round final calculations to
the precision of the least significant component. So, if somewhere in
your calculation formula you are using whole seconds (but tenths or
hundredths for other components) then final numbers should be rounded to
the nearest second. Finish times are typically taken to the nearest
whole second so that would dictate corrected times rounded to the
nearest whole second (with standard PHRF-type calculations rating are
given in whole seconds as well so that is a 2nd place where the minimal
precision is nearest whole seconds).

If virtually everyone is using some rounding convention then I would
consider that as a possible alternative to the standard conventions.
Since most scoring is rounded to the nearest whole second that would be
a 2nd reason to round to the nearest whole second.

Finally, if your scoring program only allows a certain kind of rounding
then I would simply use that. So long as all boats are scored the same
way there isn't any unfairness and there shouldn't be grounds for any
redress by competitors. From this perspective no method of rounding
could ever be "wrong" but rounding to the nearest whole second would be
BEST when finish times are recorded to the nearest whole second.

BOTTOM LINE: I think boats whose corrected time is the same when rounded
to the nearest whole second SHOULD be tied (based on whole second finish
times). However, it wouldn't break the RRS if such boats were scored as
NOT tied based on a difference in tenths of a second so long as the
corrected times of ALL boats were calculated to the same precision (all
boats in the same class for the whole regatta or series, not all boats
in other classes as well). In that regard, I think SailWave is just fine
rounding corrected times to the nearest second since I would guess that
99.999% of finish times are entered to the nearest second. That is
consistent with standard convention, usual scoring practice and
tradition. I do NOT see a problem here.

Art Engel

S. Mark Townsend wrote:

PROBLEM 1 - Corrected Time Precision

The corrected times of Rebel Yell and Grand Illusion are not being

calculated sufficient precision to resolve the corrected time tie
between the boats. In the 2005-2008 Racing Rules of Sailing rule A3 was
changed to remove to the phrase "corrected to the nearest second" so
that the individual rating rules now determine the precision of the
corrected results. By rounding the result to the nearest second a tie is
being created that should not exists. There is no method in SailWave to
specify the precision to which you want the corrected time calculated.

RRS 2009-2012

A3 STARTING TIMES AND FINISHING PLACES

The time of a boat's starting signal shall be her starting time, and

the order in which boats finish a race shall determine their finishing
places. However, when a handicap or rating system is used a boat's
corrected time shall determine her finishing place.

Mark Townsend
s_mark_townsend@hotmail.com

------------------------------------

-!- http://www.yachtsandyachting.com/ -!- http://www.sailing.org/ -!- http://www.sailwave.com/ -!- On-Line Sailwave help...http://sailwave.com/help/HTML ~ Mark Townsend's Sailwave User Guide is available from http://www.abyc.org/upload/Sailwave_ABYC_User_Guide.pdf ~ Convert to daily digest of emails send blank email to sailwave-digest@yahoogroups.com ~ To unsubscribe from the SUG please send blank email to sailwave-unsubscribe@yahoogroups.com Yahoo! Groups Links

------------------------------------

-!- http://www.yachtsandyachting.com/ -!- http://www.sailing.org/ -!- http://www.sailwave.com/ -!- On-Line Sailwave help...http://sailwave.com/help/HTML ~ Mark Townsend's Sailwave User Guide is available from http://www.abyc.org/upload/Sailwave_ABYC_User_Guide.pdf ~ Convert to daily digest of emails send blank email to sailwave-digest@yahoogroups.com ~ To unsubscribe from the SUG please send blank email to sailwave-unsubscribe@yahoogroups.com Yahoo! Groups Links

Gil_Vick · 28 June 2010 00:22

It is beyond my belief that anyone would think that times
under a second or two or three are even possible. Just go with what
the RC writes down – plus or minus a few!!!

…Gil

···

On 6/27/2010 6:53 PM, Art Engel wrote:

The
change was approved by the ISAF Racing Rules Comm. (and ISAF

Council) in 2002 to become effective with the 2005-08 rulebook. The

principal reason given was that it was more appropriate for such
matters

to be determined by the rating/handicap authority or the scorer, rather

than the racing rules. “Automatic” timing was not mentioned and

presumably was not a factor (or a only very minor factor).

There are plenty of obstacles to “automatic” timing. In recent years

(after the change in 2002) ISAF has apparently started to explore the

idea but so far as I know it hasn’t gotten beyond the “thinking” stage.

Others may have tried to make devices but I haven’t heard of any
success.

The “correct to the nearest second” concept was introduced into the

rulebook in the 2001-04 edition as part of an extensive redraft of

Appendix A that was done by one individual and adopted by ISAF (i.e.,

there was no formal submission). The idea of “corrected times” hadn’t

previously been mentioned by the racing rules at all. Within a year it

was determined that the mention of “corrected times” was a good
addition

to the racing rules but the idea of legislating precision (i.e., how to

round corrected times) probably was not. Hence, the “nearest second”

concept was only in the racing rules for 4 years (and for 2+ of those

years it was scheduled to be removed at the first opportunity).

Art Engel

Caveat - The “correct to nearest second” language was deleted as the

result of a submission from the USA. I was a member of the US Racing

Rules Committee at the time (and still am).

Philippe DE TROY wrote:

The precision to the second was removed in RRS 2009-2012 to allow
different

precision especially for developping electronic (automatic)
timing. However

I think this is not a good choice. Automatic timing is based on
on-board

GPS sensor using fixed beacons, hence very good precision. But the
sensor

is on the cebtreline of the boat and using this for timing implies
larger

changes in the RRS like definition of boat starting and finishing
a race…

Phil

----- Original Message -----

From: “Art Engel” artengel123@earthlink.net

To: sailwave@yahoogroups.com

Sent: Sunday, June 27, 2010 1:34 AM

Subject: Re: [sailwave] Problem with Corrected Time Precision

I think you might be mixing two issues - whether SailWave can
be set to

round to a precision defined by the user and what precision
SHOULD be

used. I think I might take issue with your statement "By
rounding the

result to the nearest second a tie is being created that
should not

exists [sic]."

I would agree that it might be helpful if SailWave could be
set to round

to say tenths or hundredths of a second. However, I think that
would

only rarely, if ever, be appropriate. Hence, I wouldn’t put it
high on a

priority list of features for a scoring program (or even
include it on

such a list).