[sailwave] Result culculations

Two finishers in the same class are a second apart in finishing. The leading boat has clearly

beaten the next finisher.

To me, (and I must stress I’m speaking personally) there is nothing clear about it. Handicap racing is primarily against the clock. If the boats tie on the clock then they have tied. It really doesn’t seem to me logical to do it the other way.

Lets be in a dream world where results are accurately recorded to a tenth of a second. Two boats of a handicap around 1190 can be 1.1 seconds apart on elapsed, about 0.9 seconds on corrected and still tie to rounded seconds. Some say that’s wrong, and artificially part them in the ways described above…

But consider, if there are boats with very different handicaps they could also be 0.9 seconds apart if they were recorded to tenths of a second. But, because most of us can only record times to the nearest second (at best), it means that we cannot artificially part boats that were just as far apart, but on different handicaps.

So this means that just because the two tieing boats happen to have the same handicap we are splitting them, but if the two boats were of different handicaps, even though the difference in the actual corrected times, were we able to measure them, was no less, then we wouldn’t split them. The end result, then, is that by artificially splitting boats with identical corrected times, we are introducing an inconsistency in our scoring which can quite reasonably be described as unfair.

Jim C

Jim,

The original poster hasn't responded to clarify his comment.

I suspect most of us, if not all, would agree that when two boats with
different ratings happen to tie on corrected time then there is nothing
wrong with that. I think sailors fundamentally "get" that so there
should be no perception problem in that situation.

However, when two boats that are identical and have the same rating
clearly finish one behind the other then I think sailors would be
"confused" when the results show the two boats tied. As I see it, that
is a "flaw" (or perhaps better - a "shortcoming") of the scoring system
being used (and not in the software).

REAL EXAMPLES

Two boats with TCF ratings of 0.981. The elapsed, corrected without
rounding and corrected with rounding are:

A - 25:52 - 25:22.512 - 25:23
B - 25:53 - 25:23.493 - 25:23

A - 59:12 - 58:04.512 - 58:05
B - 59:13 - 58:05.493 - 58:05

As I said earlier, personally I would add 1 second to Boat B so that
there is no tie as that would be consistent with what sailors expect
and, I think, what they want.

BIGGER ISSUE

Looked at from another perspective, you raise an interesting issue - in
an ideal world, what should the scoring be?

I believe that a statistician (I am NOT one although I started out in
college as a math major) would look at the situation as follows:

We know for a fact that Boats A and B are "equal" and that Boat A beat
Boat B "on the water." If we use only the finish/elapsed times and
calculated corrected times to score we would be throwing away real data
- in other words, we'd be ignoring that Boat A clearly finished ahead of
Boat B. So, a statistician would probably have a problem with scoring
those boats tied as the "scoring system" has ignored relevant data in
the possession of the RC.

The situation is different when the boats had different ratings and did
not finish close to each other. There the only data we have are the
finish/elapsed times and we don't have other data to use. So, in that
situation leaving a tie makes sense because we don't have the additional
data to break it. If we were to not round at all and use corrected times
to say 5 decimal places then we'd be adding information (increments of
seconds) that wasn't in the data collected by the RC. It would amount to
randomly breaking potential ties (which is not per se unfair) so it
might be hard to attack it on fairness grounds though not on some other
rational basis. Deciding who wins races by a lottery is not "unfair" but
we wouldn't accept it as it has no rational relationship to sailing
ability or skill.

I think the bottom line of this discussion is that Sailwave scores
properly and correctly but as scorers we should be reviewing results for
possible anomalies before posting. We need to be aware that ultimately
the scoring should be consistent with what sailors are expecting. Hence,
I'd add 1 second to Boat B in my examples above even though that
technically would not be strictly consistent with the rules of the
scoring system (which presumably assume that we won't randomly change
times from those recorded by the RC).

Art

···

On 12/28/2016 1:34 PM, yho@devboats.co.uk [sailwave] wrote:

Two finishers in the same class are a second apart in finishing.
The leading boat has clearly beaten the next finisher.

To me, (and I must stress I'm speaking personally) there is nothing
clear about it. Handicap racing is primarily against the clock. If
the boats tie on the clock then they have tied. It really doesn't
seem to me logical to do it the other way.

Lets be in a dream world where results are accurately recorded to a
tenth of a second. Two boats of a handicap around 1190 can be 1.1
seconds apart on elapsed, about 0.9 seconds on corrected and still
tie to rounded seconds. Some say that's wrong, and artificially part
them in the ways described above..

But consider, if there are boats with very different handicaps they
could also be 0.9 seconds apart if they were recorded to tenths of a
second. But, because most of us can only record times to the nearest
second (at best), it means that we cannot artificially part boats
that were just as far apart, but on different handicaps.

So this means that just because the two tieing boats happen to have
the same handicap we are splitting them, but if the two boats were of
different handicaps, even though the difference in the actual
corrected times, were we able to measure them, was no less, then we
wouldn't split them. The end result, then, is that by artificially
splitting boats with identical corrected times, we are introducing an
inconsistency in our scoring which can quite reasonably be described
as unfair.

Jim C

Dear all,

To update you, I sail a Laser at the local club.

There have been times when Laser’s have sailed the weekly handicap series on a Wednesday night, and Laser’s finish a second apart, yet get the same corrected finish time. On another night, the 1 second different will remain in the
corrected finish time.

As you say, the RYA rules could allow this to happen…

Regards

Pete

···

On 27 Dec 2016, at 23:14, Art Engel artengel123@earthlink.net [sailwave] sailwave@yahoogroups.com wrote:

Your situation isn’t completely clear as you left out some important facts.

If your two boats have different handicap ratings then whether or not
they are tied is real and should stand.

If your two boats have the same handicap rating AND you are using TCF
(time on time) scoring then I suppose it is possible that under the RYA
rules (which it has been reported here require rounding to the nearest
second) you could have a tie based on rounding. In theory, the RYA rules
mandate that those boats should be scored tied and it would be improper
to artificially break the tie. Personally, I would consider the RYA
rules to be “wrong and unfair” in that situation and I’d simply change
one of the times for the two boats by 1 second. No one will ever know
and the results will be consistent with what the competitors saw on the
water. However, you need to recognize that you would be technically
breaking the RYA rules and in theory a competitor could ask for redress
if they knew the underlying facts. You’ll have to decide for yourself
how you want to handle this.

This is a “flaw” inherent in TCF scoring. I’d have a problem if Sailwave
scored other than required by the RYA rules. As I said, I’d “fix” the
problem by adjusting one of the two times before publishing any results.
I wouldn’t want any scoring software to try and “fix” this kind of
problem, which comes from the handicap/rating system and NOT from the
software itself.

This points out what I think we all understand - handicap/rating systems
all have flaws and quirks that many would consider make them
“inaccurate” in the bigger picture. Such is life!

Art

PS - I suspect the RYA would completely (but possibly unofficially)
support the “fix” of adjusting one of the two times. Their rules won’t
cover it because they would consider it a “rare” event.

On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.uk [sailwave]
wrote:

Dear all,

I’ve been following this thread, and have seen it happen within our
club during handicap racing.

Two finishers in the same class are a second apart in finishing. The
leading boat has clearly beaten the next finisher.

When the numbers are put into sailwave, and the handicaps applied,
and the elapsed times are calculated according to the RYA rules,
either situation can occur:

Both boats receive the same time, and score the same points in the
race. Each boat receives a different time, with the leading boat
getting the lower points.

If you are in control of the results, and have enough time space, you
can “adjust” the times to give the correct result.

I have found this issue very frustrating on a number of occasions,
but I’m unable to control the input/output. You beat a boat on the
water, yet the results say it was a tie.

I have thought about calculating to 0.1 of a second, but this
requires a change to rules and results program.

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 07:22, ‘e-mail ian.frogley’
ian.frogley@ntlworld.commailto:ian.frogley@ntlworld.com [sailwave]
<sailwave@yahoogroups.commailto:sailwave@yahoogroups.com> wrote:

The RYA PY Scheme states “All times in seconds, any decimals are
rounded to the nearest whole figure with a decimal of 0.5 and above
rounded up and less than 0.5 down” which is patently sensible, given
that we are not able to record a finish time more accurately than to
the nearest second. Also the RYA NHC Scheme also rounds to the
nearest second “All corrected times shall be rounded to the nearest
second with 0.5 seconds being rounded up.”

As an engineer who was involved in taking measurements in all sorts
of situations I have to agree with Art that there can be no valid
reason to calculate a corrected time to fractions of a second, when
the rating of the crafts involved are averaged over a number of
similar craft, and the recorded finish time is only to the nearest
second.

IMO all rating systems should include rounding to the nearest second.
Those that do not need to be changed.

Ian Frogley

Our club uses NHC calculations. I get a lot of “feedback” about same class boats (J80s) for example, who now owe each other time. As you say, sailboat is not accurate or “fair” in everyone’s eyes.

···

On 27 Dec 2016, at 23:14, Art Engel artengel123@earthlink.net [sailwave] sailwave@yahoogroups.com wrote:

Your situation isn’t completely clear as you left out some important facts.

If your two boats have different handicap ratings then whether or not
they are tied is real and should stand.

If your two boats have the same handicap rating AND you are using TCF
(time on time) scoring then I suppose it is possible that under the RYA
rules (which it has been reported here require rounding to the nearest
second) you could have a tie based on rounding. In theory, the RYA rules
mandate that those boats should be scored tied and it would be improper
to artificially break the tie. Personally, I would consider the RYA
rules to be “wrong and unfair” in that situation and I’d simply change
one of the times for the two boats by 1 second. No one will ever know
and the results will be consistent with what the competitors saw on the
water. However, you need to recognize that you would be technically
breaking the RYA rules and in theory a competitor could ask for redress
if they knew the underlying facts. You’ll have to decide for yourself
how you want to handle this.

This is a “flaw” inherent in TCF scoring. I’d have a problem if Sailwave
scored other than required by the RYA rules. As I said, I’d “fix” the
problem by adjusting one of the two times before publishing any results.
I wouldn’t want any scoring software to try and “fix” this kind of
problem, which comes from the handicap/rating system and NOT from the
software itself.

This points out what I think we all understand - handicap/rating systems
all have flaws and quirks that many would consider make them
“inaccurate” in the bigger picture. Such is life!

Art

PS - I suspect the RYA would completely (but possibly unofficially)
support the “fix” of adjusting one of the two times. Their rules won’t
cover it because they would consider it a “rare” event.

On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.uk [sailwave]
wrote:

Dear all,

I’ve been following this thread, and have seen it happen within our
club during handicap racing.

Two finishers in the same class are a second apart in finishing. The
leading boat has clearly beaten the next finisher.

When the numbers are put into sailwave, and the handicaps applied,
and the elapsed times are calculated according to the RYA rules,
either situation can occur:

Both boats receive the same time, and score the same points in the
race. Each boat receives a different time, with the leading boat
getting the lower points.

If you are in control of the results, and have enough time space, you
can “adjust” the times to give the correct result.

I have found this issue very frustrating on a number of occasions,
but I’m unable to control the input/output. You beat a boat on the
water, yet the results say it was a tie.

I have thought about calculating to 0.1 of a second, but this
requires a change to rules and results program.

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 07:22, ‘e-mail ian.frogley’
ian.frogley@ntlworld.commailto:ian.frogley@ntlworld.com [sailwave]
<sailwave@yahoogroups.commailto:sailwave@yahoogroups.com> wrote:

The RYA PY Scheme states “All times in seconds, any decimals are
rounded to the nearest whole figure with a decimal of 0.5 and above
rounded up and less than 0.5 down” which is patently sensible, given
that we are not able to record a finish time more accurately than to
the nearest second. Also the RYA NHC Scheme also rounds to the
nearest second “All corrected times shall be rounded to the nearest
second with 0.5 seconds being rounded up.”

As an engineer who was involved in taking measurements in all sorts
of situations I have to agree with Art that there can be no valid
reason to calculate a corrected time to fractions of a second, when
the rating of the crafts involved are averaged over a number of
similar craft, and the recorded finish time is only to the nearest
second.

IMO all rating systems should include rounding to the nearest second.
Those that do not need to be changed.

Ian Frogley

Steve,

A reminder that the RYA "positively encourages" dual scoring with both
regular and NHC handicap ratings. Other personal handicap type systems
actually require dual scoring. I'd highly recommend it as dual scoring
highlights the fact that a personal handicap system is not designed to
reward the best sailors but rather to give every boat a chance over time
to get a 1st-place trophy, regardless of skill level.

Dual scoring usually eliminates (or reduces) the "feedback" as those
sailors who think the best sailor should win look at the regular results
and those sailors who think everyone is entitled to a 1st-place trophy
look at the personal handicap results. Everyone is equally happy (or
equally unhappy!).

Art

···

On 12/28/2016 4:09 PM, Steve Dahlstrom sdahlstrom@ymail.com [sailwave] wrote:

Our club uses NHC calculations. I get a lot of "feedback" about same
class boats (J80s) for example, who now owe each other time. As you
say, sailboat is not accurate or "fair" in everyone's eyes.

Sent from my iPad

On Dec 28, 2016, at 5:42 PM, pete scrutton pete141820@hotmail.co.uk >> [sailwave] <sailwave@yahoogroups.com> wrote:

Dear all,

To update you, I sail a Laser at the local club.

There have been times when Laser's have sailed the weekly handicap
series on a Wednesday night, and Laser's finish a second apart, yet
get the same corrected finish time. On another night, the 1 second
different will remain in the corrected finish time.

As you say, the RYA rules could allow this to happen....

Regards

Pete

Pete,

You don't say what handicapping system is being used. Assuming it is the
RYA Portsmouth Yardstick system then in essence the Portsmouth Yardstick
numbers are converted into TCFs by the formula 1000/PY (the formula for
scoring is CT=ET*(1000/PY) where ET is elapsed time and CT is corrected
time). Thus, the "inaccuracies" inherent in a TCF system exist to
exactly the same extent in the RYA's PY system.

Thanks for clarifying - it turns out to be what we have all been
assuming so our analysis and comments are directly relevant.

I haven't searched the RYA website nor contacted them but I doubt they'd
be terribly upset if you adjusted times to break ties when one boat
clearly beat another on the water. But, as it would be a rare occurrence
their rules probably don't cover it.

Art

PS - The 2016 Laser PY of 1095 is equal to a TCF of 0.913242009. The
only difference between this and a standard TCF issued under ORC is the
number of decimal places (ORC being limited to 3, I believe), which
isn't significant since the additional decimal places impact corrected
times at the level of tenths of a second.

···

On 12/28/2016 3:42 PM, pete scrutton pete141820@hotmail.co.uk [sailwave] wrote:

Dear all,

To update you, I sail a Laser at the local club.

There have been times when Laser's have sailed the weekly handicap
series on a Wednesday night, and Laser's finish a second apart, yet
get the same corrected finish time. On another night, the 1 second
different will remain in the corrected finish time.

As you say, the RYA rules could allow this to happen....

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 23:14, Art Engel > artengel123@earthlink.net<mailto:artengel123@earthlink.net> > [sailwave] > <sailwave@yahoogroups.com<mailto:sailwave@yahoogroups.com>> wrote:

Your situation isn't completely clear as you left out some important facts.

If your two boats have different handicap ratings then whether or not
they are tied is real and should stand.

If your two boats have the same handicap rating AND you are using TCF
(time on time) scoring then I suppose it is possible that under the RYA
rules (which it has been reported here require rounding to the nearest
second) you could have a tie based on rounding. In theory, the RYA rules
mandate that those boats should be scored tied and it would be improper
to artificially break the tie. Personally, I would consider the RYA
rules to be "wrong and unfair" in that situation and I'd simply change
one of the times for the two boats by 1 second. No one will ever know
and the results will be consistent with what the competitors saw on the
water. However, you need to recognize that you would be technically
breaking the RYA rules and in theory a competitor could ask for redress
if they knew the underlying facts. You'll have to decide for yourself
how you want to handle this.

This is a "flaw" inherent in TCF scoring. I'd have a problem if Sailwave
scored other than required by the RYA rules. As I said, I'd "fix" the
problem by adjusting one of the two times before publishing any results.
I wouldn't want any scoring software to try and "fix" this kind of
problem, which comes from the handicap/rating system and NOT from the
software itself.

This points out what I think we all understand - handicap/rating systems
all have flaws and quirks that many would consider make them
"inaccurate" in the bigger picture. Such is life!

Art

PS - I suspect the RYA would completely (but possibly unofficially)
support the "fix" of adjusting one of the two times. Their rules won't
cover it because they would consider it a "rare" event.

On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.uk<mailto:pete141820@hotmail.co.uk> [sailwave] > wrote:

Dear all,

I've been following this thread, and have seen it happen within our
club during handicap racing.

Two finishers in the same class are a second apart in finishing. The
leading boat has clearly beaten the next finisher.

When the numbers are put into sailwave, and the handicaps applied,
and the elapsed times are calculated according to the RYA rules,
either situation can occur:

Both boats receive the same time, and score the same points in the
race. Each boat receives a different time, with the leading boat
getting the lower points.

If you are in control of the results, and have enough time space, you
can "adjust" the times to give the correct result.

I have found this issue very frustrating on a number of occasions,
but I'm unable to control the input/output. You beat a boat on the
water, yet the results say it was a tie.

I have thought about calculating to 0.1 of a second, but this
requires a change to rules and results program.

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 07:22, 'e-mail ian.frogley' >> ian.frogley@ntlworld.com<mailto:ian.frogley@ntlworld.com><mailto:ian.frogley@ntlworld.com> [sailwave] >> <sailwave@yahoogroups.com<mailto:sailwave@yahoogroups.com><mailto:sailwave@yahoogroups.com>> wrote:

The RYA PY Scheme states "All times in seconds, any decimals are
rounded to the nearest whole figure with a decimal of 0.5 and above
rounded up and less than 0.5 down" which is patently sensible, given
that we are not able to record a finish time more accurately than to
the nearest second. Also the RYA NHC Scheme also rounds to the
nearest second "All corrected times shall be rounded to the nearest
second with 0.5 seconds being rounded up."

As an engineer who was involved in taking measurements in all sorts
of situations I have to agree with Art that there can be no valid
reason to calculate a corrected time to fractions of a second, when
the rating of the crafts involved are averaged over a number of
similar craft, and the recorded finish time is only to the nearest
second.

IMO all rating systems should include rounding to the nearest second.
Those that do not need to be changed.

Ian Frogley

Trust me, they all have their own calculations to base the beer at the club on…

···

Sent from my iPad

On Dec 28, 2016, at 6:32 PM, Art Engel artengel123@earthlink.net [sailwave] sailwave@yahoogroups.com wrote:

Pete,

You don’t say what handicapping system is being used. Assuming it is the
RYA Portsmouth Yardstick system then in essence the Portsmouth Yardstick
numbers are converted into TCFs by the formula 1000/PY (the formula for
scoring is CT=ET*(1000/PY) where ET is elapsed time and CT is corrected
time). Thus, the “inaccuracies” inherent in a TCF system exist to
exactly the same extent in the RYA’s PY system.

Thanks for clarifying - it turns out to be what we have all been
assuming so our analysis and comments are directly relevant.

I haven’t searched the RYA website nor contacted them but I doubt they’d
be terribly upset if you adjusted times to break ties when one boat
clearly beat another on the water. But, as it would be a rare occurrence
their rules probably don’t cover it.

Art

PS - The 2016 Laser PY of 1095 is equal to a TCF of 0.913242009. The
only difference between this and a standard TCF issued under ORC is the
number of decimal places (ORC being limited to 3, I believe), which
isn’t significant since the additional decimal places impact corrected
times at the level of tenths of a second.

On 12/28/2016 3:42 PM, pete scrutton pete141820@hotmail.co.uk [sailwave] > wrote:

Dear all,

To update you, I sail a Laser at the local club.

There have been times when Laser’s have sailed the weekly handicap
series on a Wednesday night, and Laser’s finish a second apart, yet
get the same corrected finish time. On another night, the 1 second
different will remain in the corrected finish time.

As you say, the RYA rules could allow this to happen…

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 23:14, Art Engel > > artengel123@earthlink.netmailto:artengel123@earthlink.net > > [sailwave] > > <sailwave@yahoogroups.commailto:sailwave@yahoogroups.com> wrote:

Your situation isn’t completely clear as you left out some important facts.

If your two boats have different handicap ratings then whether or not
they are tied is real and should stand.

If your two boats have the same handicap rating AND you are using TCF
(time on time) scoring then I suppose it is possible that under the RYA
rules (which it has been reported here require rounding to the nearest
second) you could have a tie based on rounding. In theory, the RYA rules
mandate that those boats should be scored tied and it would be improper
to artificially break the tie. Personally, I would consider the RYA
rules to be “wrong and unfair” in that situation and I’d simply change
one of the times for the two boats by 1 second. No one will ever know
and the results will be consistent with what the competitors saw on the
water. However, you need to recognize that you would be technically
breaking the RYA rules and in theory a competitor could ask for redress
if they knew the underlying facts. You’ll have to decide for yourself
how you want to handle this.

This is a “flaw” inherent in TCF scoring. I’d have a problem if Sailwave
scored other than required by the RYA rules. As I said, I’d “fix” the
problem by adjusting one of the two times before publishing any results.
I wouldn’t want any scoring software to try and “fix” this kind of
problem, which comes from the handicap/rating system and NOT from the
software itself.

This points out what I think we all understand - handicap/rating systems
all have flaws and quirks that many would consider make them
“inaccurate” in the bigger picture. Such is life!

Art

PS - I suspect the RYA would completely (but possibly unofficially)
support the “fix” of adjusting one of the two times. Their rules won’t
cover it because they would consider it a “rare” event.

On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.ukmailto:pete141820@hotmail.co.uk [sailwave] > > wrote:

Dear all,

I’ve been following this thread, and have seen it happen within our
club during handicap racing.

Two finishers in the same class are a second apart in finishing. The
leading boat has clearly beaten the next finisher.

When the numbers are put into sailwave, and the handicaps applied,
and the elapsed times are calculated according to the RYA rules,
either situation can occur:

Both boats receive the same time, and score the same points in the
race. Each boat receives a different time, with the leading boat
getting the lower points.

If you are in control of the results, and have enough time space, you
can “adjust” the times to give the correct result.

I have found this issue very frustrating on a number of occasions,
but I’m unable to control the input/output. You beat a boat on the
water, yet the results say it was a tie.

I have thought about calculating to 0.1 of a second, but this
requires a change to rules and results program.

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 07:22, ‘e-mail ian.frogley’ > >> ian.frogley@ntlworld.commailto:ian.frogley@ntlworld.commailto:ian.frogley@ntlworld.com [sailwave] > >> <sailwave@yahoogroups.commailto:sailwave@yahoogroups.commailto:sailwave@yahoogroups.com> wrote:

The RYA PY Scheme states “All times in seconds, any decimals are
rounded to the nearest whole figure with a decimal of 0.5 and above
rounded up and less than 0.5 down” which is patently sensible, given
that we are not able to record a finish time more accurately than to
the nearest second. Also the RYA NHC Scheme also rounds to the
nearest second “All corrected times shall be rounded to the nearest
second with 0.5 seconds being rounded up.”

As an engineer who was involved in taking measurements in all sorts
of situations I have to agree with Art that there can be no valid
reason to calculate a corrected time to fractions of a second, when
the rating of the crafts involved are averaged over a number of
similar craft, and the recorded finish time is only to the nearest
second.

IMO all rating systems should include rounding to the nearest second.
Those that do not need to be changed.

Ian Frogley

I don’t think it is an exact parallel but try this for an
anomaly: let us assume that two’Laser’ 2000s with a PN of 1100
have each done three laps and both have an elapsed time of 3000
seconds. Elsewhere in the fleet a Musto Skiff has done five laps.
The club race box contains an easy to use calculator with large
buttons but it is designed for shop use and does not display
decimals (although it may store them during long calculations). I
look at the first 2000’s ET and, coming from a generation which
did not have calculators at its disposal, I mentally go 3000
divided by 3 is a thousand, times five is 5000 times ten is fifty
thousand divided by 11 is 4545 seconds (looks about right). For
the second 2000 I use the club’s calculator as follows: 5 divided
by 3 equals ? times 3000 times1000 divided by 1100 equals 5454
seconds. Looks the same as the other boat, continue to next
boat… And yet the two corrected times are different by
almost 900 seconds! The fault stems from my first ‘equals’ which
should have produced a figure of 1.6666667 but the calculator has
rounded it up to 2 which I haven’t noticed and I have continued
pressing buttons to produce the fluke result which contains
exactly the same numbers but in a different order, which I don’t
notice. This example is contrived but I believe that rounding
errors do occur if you do calculations with the same data in
different ways. But your generation knows that - I shall have to
revert to logarithms!

rgds

George Morris

···

On 29/12/2016 00:09, Steve Dahlstrom
[sailwave] wrote:

sdahlstrom@ymail.com

    Our club uses NHC calculations. I get a lot of "feedback"

about same class boats (J80s) for example, who now owe each
other time. As you say, sailboat is not accurate or “fair” in
everyone’s eyes.

    Sent from my iPad
    On Dec 28, 2016, at 5:42 PM, pete scrutton pete141820@hotmail.co.uk
    [sailwave] <sailwave@yahoogroups.com        >

wrote:

Dear all,

          To update you, I sail a Laser

at the local club.

          There have been times when

Laser’s have sailed the weekly handicap series on a
Wednesday night, and Laser’s finish a second apart, yet
get the same corrected finish time. On another night, the
1 second different will remain in the corrected finish
time.

          As you say, the RYA rules could

allow this to happen…

          Regards

Pete

          Sent from my iPad
          On 27 Dec 2016, at 23:14, Art Engel artengel123@earthlink.net
          [sailwave] <sailwave@yahoogroups.com              >

wrote:

                Your situation isn't completely clear as you left

out some important facts.

                If your two boats have different handicap ratings

then whether or not
they are tied is real and should stand.

                If your two boats have the same handicap rating AND

you are using TCF
(time on time) scoring then I suppose it is possible
that under the RYA
rules (which it has been reported here require
rounding to the nearest
second) you could have a tie based on rounding. In
theory, the RYA rules
mandate that those boats should be scored tied and
it would be improper
to artificially break the tie. Personally, I would
consider the RYA
rules to be “wrong and unfair” in that situation and
I’d simply change
one of the times for the two boats by 1 second. No
one will ever know
and the results will be consistent with what the
competitors saw on the
water. However, you need to recognize that you would
be technically
breaking the RYA rules and in theory a competitor
could ask for redress
if they knew the underlying facts. You’ll have to
decide for yourself
how you want to handle this.

                This is a "flaw" inherent in TCF scoring. I'd have a

problem if Sailwave
scored other than required by the RYA rules. As I
said, I’d “fix” the
problem by adjusting one of the two times before
publishing any results.
I wouldn’t want any scoring software to try and
“fix” this kind of
problem, which comes from the handicap/rating system
and NOT from the
software itself.

                This points out what I think we all understand -

handicap/rating systems
all have flaws and quirks that many would consider
make them
“inaccurate” in the bigger picture. Such is life!

                Art

                PS - I suspect the RYA would completely (but

possibly unofficially)
support the “fix” of adjusting one of the two times.
Their rules won’t
cover it because they would consider it a “rare”
event.

                On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.uk
                [sailwave]
                wrote:
                > Dear all,
                >
                > I've been following this thread, and have seen

it happen within our
> club during handicap racing.
>
> Two finishers in the same class are a second
apart in finishing. The
> leading boat has clearly beaten the next
finisher.
>
> When the numbers are put into sailwave, and the
handicaps applied,
> and the elapsed times are calculated according
to the RYA rules,
> either situation can occur:
>
> Both boats receive the same time, and score the
same points in the
> race. Each boat receives a different time, with
the leading boat
> getting the lower points.
>
> If you are in control of the results, and have
enough time space, you
> can “adjust” the times to give the correct
result.
>
> I have found this issue very frustrating on a
number of occasions,
> but I’m unable to control the input/output. You
beat a boat on the
> water, yet the results say it was a tie.
>
> I have thought about calculating to 0.1 of a
second, but this
> requires a change to rules and results program.
>
> Regards
>
> Pete
>
>
> Sent from my iPad
>
> On 27 Dec 2016, at 07:22, ‘e-mail ian.frogley’
> ian.frogley@ntlworld.com<mailto:ian.frogley@ntlworld.com >
[sailwave]
> <sailwave@yahoogroups.com<mailto:sailwave@yahoogroups.com >>
wrote:
>
>
> The RYA PY Scheme states “All times in seconds,
any decimals are
> rounded to the nearest whole figure with a
decimal of 0.5 and above
> rounded up and less than 0.5 down” which is
patently sensible, given
> that we are not able to record a finish time
more accurately than to
> the nearest second. Also the RYA NHC Scheme
also rounds to the
> nearest second “All corrected times shall be
rounded to the nearest
> second with 0.5 seconds being rounded up.”
>
> As an engineer who was involved in taking
measurements in all sorts
> of situations I have to agree with Art that
there can be no valid
> reason to calculate a corrected time to
fractions of a second, when
> the rating of the crafts involved are averaged
over a number of
> similar craft, and the recorded finish time is
only to the nearest
> second.
>
> IMO all rating systems should include rounding
to the nearest second.
> Those that do not need to be changed.
>
> Ian Frogley
>

Rule A3 no longer tells you to round corrected times after applying the handicap. The rounding of corrected times is a matter for the handicap or rating system to specify. IRC and RYA Portsmouth Yardstick, for example, have a time-rounding rule, PHRF
does not.

The following is an example of why rounding could be considered unfair.

A PHRF race being scored using time on distance handicap.

Race distance 3.5 miles

Boat A - Handicap 3

Adjustment of 3 seconds per mile. 3.5 miles x 3 secs = 10.5 seconds

Elapsed time 40 min 11 seconds

Handicap Adjustment -10.5

Corrected time 40 min 0.5 seconds

Boat B - Handicap 6

Adjustment of 6 seconds per mile. 3.5 miles x 6 secs = 21 seconds

Elapsed time 40 min 21 seconds

Handicap Adjustment -21

Corrected time 40 min 0 seconds

If you round Boat A’s corrected time you are not giving them the benefit of their full handicap. That would seem to be unfair.

···

Mark Townsend

Phone: 562-433-4366

Mobile: 562-533-5909

Email:
s_mark_townsend@hotmail.com


From: sailwave@yahoogroups.com sailwave@yahoogroups.com on behalf of George Morris georgewhitleymorris@gmail.com [sailwave] sailwave@yahoogroups.com
Sent: Wednesday, December 28, 2016 5:20 PM
To: sailwave@yahoogroups.com
Subject: Re: [sailwave] Result culculations

I don’t think it is an exact parallel but try this for an anomaly: let us assume that two’Laser’ 2000s with a PN of 1100 have each done three laps and both have an elapsed time of 3000 seconds. Elsewhere in the fleet a Musto Skiff has done five laps. The
club race box contains an easy to use calculator with large buttons but it is designed for shop use and does not display decimals (although it may store them during long calculations). I look at the first 2000’s ET and, coming from a generation which did not
have calculators at its disposal, I mentally go 3000 divided by 3 is a thousand, times five is 5000 times ten is fifty thousand divided by 11 is 4545 seconds (looks about right). For the second 2000 I use the club’s calculator as follows: 5 divided by 3 equals
? times 3000 times1000 divided by 1100 equals 5454 seconds. Looks the same as the other boat, continue to next boat… And yet the two corrected times are different by almost 900 seconds! The fault stems from my first ‘equals’ which should have produced
a figure of 1.6666667 but the calculator has rounded it up to 2 which I haven’t noticed and I have continued pressing buttons to produce the fluke result which contains exactly the same numbers but in a different order, which I don’t notice. This example is
contrived but I believe that rounding errors do occur if you do calculations with the same data in different ways. But your generation knows that - I shall have to revert to logarithms!

rgds

George Morris

On 29/12/2016 00:09, Steve Dahlstrom
sdahlstrom@ymail.com [sailwave] wrote:

Our club uses NHC calculations. I get a lot of “feedback” about same class boats (J80s) for example, who now owe each other time. As you say, sailboat is not accurate or “fair” in everyone’s eyes.

Sent from my iPad

On Dec 28, 2016, at 5:42 PM, pete scrutton
pete141820@hotmail.co.uk [sailwave] sailwave@yahoogroups.com wrote:

Dear all,

To update you, I sail a Laser at the local club.

There have been times when Laser’s have sailed the weekly handicap series on a Wednesday night, and Laser’s finish a second apart, yet get the same corrected finish time. On another night, the 1 second different will remain in the
corrected finish time.

As you say, the RYA rules could allow this to happen…

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 23:14, Art Engel artengel123@earthlink.net [sailwave] sailwave@yahoogroups.com wrote:

Your situation isn’t completely clear as you left out some important facts.

If your two boats have different handicap ratings then whether or not
they are tied is real and should stand.

If your two boats have the same handicap rating AND you are using TCF
(time on time) scoring then I suppose it is possible that under the RYA
rules (which it has been reported here require rounding to the nearest
second) you could have a tie based on rounding. In theory, the RYA rules
mandate that those boats should be scored tied and it would be improper
to artificially break the tie. Personally, I would consider the RYA
rules to be “wrong and unfair” in that situation and I’d simply change
one of the times for the two boats by 1 second. No one will ever know
and the results will be consistent with what the competitors saw on the
water. However, you need to recognize that you would be technically
breaking the RYA rules and in theory a competitor could ask for redress
if they knew the underlying facts. You’ll have to decide for yourself
how you want to handle this.

This is a “flaw” inherent in TCF scoring. I’d have a problem if Sailwave
scored other than required by the RYA rules. As I said, I’d “fix” the
problem by adjusting one of the two times before publishing any results.
I wouldn’t want any scoring software to try and “fix” this kind of
problem, which comes from the handicap/rating system and NOT from the
software itself.

This points out what I think we all understand - handicap/rating systems
all have flaws and quirks that many would consider make them
“inaccurate” in the bigger picture. Such is life!

Art

PS - I suspect the RYA would completely (but possibly unofficially)
support the “fix” of adjusting one of the two times. Their rules won’t
cover it because they would consider it a “rare” event.

On 12/27/2016 2:29 PM, pete scrutton pete141820@hotmail.co.uk [sailwave]
wrote:

Dear all,

I’ve been following this thread, and have seen it happen within our
club during handicap racing.

Two finishers in the same class are a second apart in finishing. The
leading boat has clearly beaten the next finisher.

When the numbers are put into sailwave, and the handicaps applied,
and the elapsed times are calculated according to the RYA rules,
either situation can occur:

Both boats receive the same time, and score the same points in the
race. Each boat receives a different time, with the leading boat
getting the lower points.

If you are in control of the results, and have enough time space, you
can “adjust” the times to give the correct result.

I have found this issue very frustrating on a number of occasions,
but I’m unable to control the input/output. You beat a boat on the
water, yet the results say it was a tie.

I have thought about calculating to 0.1 of a second, but this
requires a change to rules and results program.

Regards

Pete

Sent from my iPad

On 27 Dec 2016, at 07:22, ‘e-mail ian.frogley’
ian.frogley@ntlworld.commailto:ian.frogley@ntlworld.com [sailwave]
<sailwave@yahoogroups.commailto:sailwave@yahoogroups.com> wrote:

The RYA PY Scheme states “All times in seconds, any decimals are
rounded to the nearest whole figure with a decimal of 0.5 and above
rounded up and less than 0.5 down” which is patently sensible, given
that we are not able to record a finish time more accurately than to
the nearest second. Also the RYA NHC Scheme also rounds to the
nearest second “All corrected times shall be rounded to the nearest
second with 0.5 seconds being rounded up.”

As an engineer who was involved in taking measurements in all sorts
of situations I have to agree with Art that there can be no valid
reason to calculate a corrected time to fractions of a second, when
the rating of the crafts involved are averaged over a number of
similar craft, and the recorded finish time is only to the nearest
second.

IMO all rating systems should include rounding to the nearest second.
Those that do not need to be changed.

Ian Frogley

—In sailwave@yahoogroups.com, <s_mark_townsend@…> wrote :

If you round Boat A’s corrected time you are not giving them the benefit of their

full handicap. That would seem to be unfair.

One of the things that bedevils this sort of discussion is that folk assume that the time that’s written down on the piece of paper is exact, and when recording to the nearest second Boat A finished in 2935.0 seconds and Boat B finished in 2674.0 seconds etc. But if you think about it that’s not actually the case. Depending on the timepiece in use Boat A finished somewhere between 2935.0 and 2935.9 seconds, or possibly between 2934.5 and 2935.4 seconds.

Or, if I was being honest about the difficulties of switching from looking at the line 100 yards away to the watch 6 inches away, I probably shouldn’t be any more dogmatic than to say Boat A probably finished somewhere in the range from 2934.something to 2936.something seconds.

For convenience we take one number and say boat A finished in 2935 seconds and calculate a corrected time from that, but if we were being strictly accurate we should say that boat As corrected time was between n1 and n2, but we have absolutely no idea where it fell within that range. Another boat will have time range n3 to n4, and if that range overlaps to any extent at all with n1 to n2 then we have no idea which one really had the lowest corrected time, so we should call it a tie.

By rounding to the nearest second we approximate to this in a way that saves an awful lot of calculation, not to mention the mind blowing concept that Boat Bs range of possible times may overlap with both Boat As and Boat Cs, but A and Cs possible times don’t overlap! Its the art of the practical and possible…

Personally I’ve never found that sailors at my club have any problem at all with the concept that yes, Fred was just ahead on the water, but not far enough ahead to get a separate time on handicap, so they’re scored as a tie. To be honest the thought of fabricating false data to break a tie artificially rather horrified me.

Oh, and George, if your club’s calculator really behaves as described and doesn’t internally store the decimals then run, don’t walk to the clubhouse, throw it on the ground, stamp on it and grind the pieces into the dust!

Jim C

(again please note these opinions are entirely personal, and should not be associated in any way with the RYA EHAG).

Jim,

This has been an interesting discussion. I agree with 99.9% of what you
say but differ very slightly in my view of one situation (which is
similar to that described by the original poster).

We rarely use TCF scoring locally so we haven't had the case of
different finish times for two boats with identical ratings calculating
as a tie. However, we often run mixed handicap and OD events where we
have to use scoring by finishing times for all boats (since it isn't
always possible to know whether a particular finisher is sailing in a OD
class or a handicap class). In that case, two OD boats might finish
overlapped with one clearly crossing the finishing line ahead of the
other. In some cases, our timing mechanism records identical times for
those boats because it only records to whole seconds.

When scoring by finishing times those boats show as tied. But, if we
scored by finishing places they'd be 1-2. I don't believe the results of
a race should vary depending on the method of recording finishes.

Personally I've never found that sailors at my club have any problem
at all with the concept that yes, Fred was just ahead on the water,
but not far enough ahead to get a separate time on handicap, so
they're scored as a tie. To be honest the thought of fabricating
false data to break a tie artificially rather horrified me.

I agree that in general sailors understand "he corrected out on me."
They may not like it but they understand it. However, in the situation
I've described we've pretty much always had exactly the opposite
reaction from sailors - "How can you score us as tied when we clearly
beat them?" In that case we change the time of one of the two boats by 1
second to make the results more accurate, and because this issue has
occasionally arisen in the past we try to do it before any results are
posted so it never becomes a talking point and the sailors see exactly
what they would expect.

I'd say we are artificially changing "times" but I wouldn't say we are
artificially changing "data" since part of the data we have to work with
is the order of finish. As I said, I don't think results should change
based on the method of recording finishes used by the RC.

The situation posed by the original poster was similar although slightly
different - two OD-type boats sailing in a handicap class rather then in
a separate OD class. Because we use time-on-distance scoring locally I
don't think that exact situation has ever arisen but if it did I think
I'd handle it the same. To my way of thinking, I am putting back into
the finishing data information that was removed by our timing mechanism,
which can only record in whole seconds (for the various reasons you so
describe so well).

Art

PS - At this point, I doubt the conversation is doing much to educate or
inform its participants. But, I am responding in detail so that when
people find it in the future by searching it will be more complete.

PPS - I don't usually bother to give this caveat but as Jim says - my
opinions are my own and don't represent the official view of the US
Sailing Rules Committee or US Sailing.

···

On 12/29/2016 6:04 AM, yho@devboats.co.uk [sailwave] wrote:

---In sailwave@yahoogroups.com, <s_mark_townsend@...> wrote :

If you round Boat A's corrected time you are not giving them the
benefit of their full handicap. That would seem to be unfair.

One of the things that bedevils this sort of discussion is that folk
assume that the time that's written down on the piece of paper is
exact, and when recording to the nearest second Boat A finished in
2935.0 seconds and Boat B finished in 2674.0 seconds etc. But if you
think about it that's not actually the case. Depending on the
timepiece in use Boat A finished somewhere between 2935.0 and 2935.9
seconds, or possibly between 2934.5 and 2935.4 seconds.

Or, if I was being honest about the difficulties of switching from
looking at the line 100 yards away to the watch 6 inches away, I
probably shouldn't be any more dogmatic than to say Boat A probably
finished somewhere in the range from 2934.something to 2936.something
seconds.

For convenience we take one number and say boat A finished in 2935
seconds and calculate a corrected time from that, but if we were
being strictly accurate we should say that boat As corrected time was
between n1 and n2, but we have absolutely no idea where it fell
within that range. Another boat will have time range n3 to n4, and
if that range overlaps to any extent at all with n1 to n2 then we
have no idea which one really had the lowest corrected time, so we
should call it a tie.

By rounding to the nearest second we approximate to this in a way
that saves an awful lot of calculation, not to mention the mind
blowing concept that Boat Bs range of possible times may overlap with
both Boat As and Boat Cs, but A and Cs possible times don't overlap!
Its the art of the practical and possible...

Personally I've never found that sailors at my club have any problem
at all with the concept that yes, Fred was just ahead on the water,
but not far enough ahead to get a separate time on handicap, so
they're scored as a tie. To be honest the thought of fabricating
false data to break a tie artificially rather horrified me.

Oh, and George, if your club's calculator really behaves as described
and doesn't internally store the decimals then run, don't walk to the
clubhouse, throw it on the ground, stamp on it and grind the pieces
into the dust!

Jim C

(again please note these opinions are entirely personal, and should
not be associated in any way with the RYA EHAG).

Interesting, I can see where you’re coming from, but the concept still rather repels me!

We have one club series where the results are scored 3 different ways:

  1. Standard handicap per class, scored to nearest (rounded) second

  2. personal handicap, also to nearest second, but handicaps based on sailor’s past performance, not just boat class

  3. places (for larger classes only), as on the water.

Its honestly never really occurred to me it might be thought a significant problem that a boat can tie on handicap and not on places. Always seemed to me more important to treat all ties in each series consistently, whether they are between boats of the same class or boats of quite different performance. But either way you end up with inconsistent treatment.

Interesting that only in the course of this dialog have I’ve realised that an automated system which records times to the nearest tenth of a second and rounds the numbers will give getting on for half of the times 1 second lower than the manual technique of noting what integer a digital watch displays as the line is crossed

JimC

The "problems" each of us chooses to deal with are the ones we see.

I've never seen two boats with the same handicap rating but different
finish times end up as tied in the results - but that is primarily
because that is a shortcoming of a system we don't use to score with
locally - the TCF-type system (corrected times derived by multiplying
elapsed times by some multiplier). If it arose I'd be inclined to "fix"
it by adjusting one of the times so the results are consistent with the
order of finishing places. I think that would be fair but then leaving
the tie would in my view be fair as well. As an active competitor
(coming late in life to RC work) I prefer the "fix" as the better of two
equally fair alternatives.

I have seen times from a finishing software program come out as the same
when one boat was clearly in front of another. When the boats have
different ratings it isn't a problem so nothing is done about it. When
the boats have the same rating it is a problem so we address it by
adjusting times. I'm not happy changing times but I much prefer that to
having the results show boats as tied when virtually no one thinks they
should be scored that way.

In our season-long series we have some 1800 boats scored. Each year
there are a handful of ties on corrected time between boats with
different ratings and we've never heard a peep from anyone (that I know
of anyway). When we have boats with the same rating tied in the
corrected results when one finished ahead of the other on the water we
ALWAYS hear about it and "fix" it. In 20+ years of protest/RC activities
I think I've only seen an "on the water" tie once - when the PRO said
the boats were absolutely bow to bow and there was no way to say one
finished ahead of the other.

Your mileage may vary!

If you don't have the problem we have then our solution may seem strange
to you but if you had the problem on a somewhat regular basis I'm fairly
confident you'd be applying the same solution.

By the way, our RC records three pieces of data for finishers: (1) sail
number, (2) finish order and (3) finish times. We input the finish times
into our scoring program (Sailwave for weekend events) but always have
the finish order as back up. When boats are tied on the water (happened
once in 20+ years that I can recall) they get the same finishing order
position.

Art

···

On 12/29/2016 4:37 PM, yho@devboats.co.uk [sailwave] wrote:

Interesting, I can see where you're coming from, but the concept
still rather repels me!

We have one club series where the results are scored 3 different
ways:

1) Standard handicap per class, scored to nearest (rounded) second

2) personal handicap, also to nearest second, but handicaps based on
sailor's past performance, not just boat class

3) places (for larger classes only), as on the water.

Its honestly never really occurred to me it might be thought a
significant problem that a boat can tie on handicap and not on
places. Always seemed to me more important to treat all ties in each
series consistently, whether they are between boats of the same class
or boats of quite different performance. But either way you end up
with inconsistent treatment.

Interesting that only in the course of this dialog have I've realised
that an automated system which records times to the nearest tenth of
a second and rounds the numbers will give getting on for half of the
times 1 second lower than the manual technique of noting what integer
a digital watch displays as the line is crossed

JimC