solosailor
New member
A faster rated boat (Ventus) that is 40nm behind a slower rated boat (Kato) shouldn't show 13 hours ahead correct should it?
-
Ahoy and Welcome to the New SSS Forums!!
As you can see, we have migrated our old forums to new software. All your old posts, threads, attachments, and messages should be here. If you see anything out of place or have any questions, please scroll to the very bottom of the page and click "Contact Us" and leave a note with as much detail as possible.
You should be able to login with your old credentials. If you have any issues, try resetting your password before clicking the Contact Us link.
Cheers - Bryan
Leaderboard?
- Thread starter solosailor
- Start date
Philpott
Cal 2-27 Dura Mater
Be advised that there is a flurry of activity behind the scenes and the crack team RC is cogitating about that issue as you type. Christine Weaver, Jonathan Gutoff and I arrived at Lihue yesterday within minutes of each other. They flew Hawaiian air and got a complimentary breakfast and rum drinks. I flew United and got very bad coffee and one cookie. I will post RC photos here because this is the SSS site, right? This site will remain long after Word Press gets sold to Tesla or Credite Dauphin Internationale.
Wish you were here, Greg and Bob. Wish you were all here! This is fun!
Last edited:
Philpott
Cal 2-27 Dura Mater
Thanks, Lucie. Chris and I are going to go to the botanical gardens and snorkeling. We probably won't go snorkeling at the botanical gardens but I am open to all experiences Kauain so if that's what is done then I will try it. Also surfing.
Last edited:
Let me know if you plan to go surfing and want to rent a board and lessons. There's a guy I met down there 4 years ago that sets up near one of the hotels. I can do some searching and figure out where, if you're interested. Have fun snorkeling, but you might want to time it for when the tide is coming in.
BobJ
Alerion 38 "Surprise!"
I can't remember what I gave Yellow Brick in 2012 when they had it screwed up. Difference in ratings x .589 and then average speed over the last 24 hours. Something like that.
Philpott
Cal 2-27 Dura Mater
The Question: Following Kato which seems to be many nm ahead of Ventus, which is same yacht class, but Ventus is listed as first. *Why is that?
The answerHEre is the issues:
1.* There is noise in the speed data.
2.* In essence timing is synchronized because the time to finish is based on the distance to the finish and then that time is added to the time on course.* So if two readings are 4 hours apart for X and Y, and Y was the latest entry, her distance to the island would be closer, that would be offset by her additional 4 hours on the course when the correction factor for rating is tossed into the calculation.** If we wanted this to be very non noisy we should have averaged more readings, but then you would still be wrong.* Boats that go stuck in the hole, for example, would look even worse than they do.
The prediction algorithm is fun because it is like discussing who is the best pitcher ever in baseball.Hello James,
The leader board compares the Corrected Projected Elapsed Time (CPET) of all boats in the race.*
CPET =* ((DTF/ProjectedSpeed) + TimeOnCourse) -* Rating*CourseDistance***** all adjusted to hours in this case.
The ProjectedSpeed is where the variability occurs. We have to select an algorithm to guess speed. If you average many readings of speed the results are less noisy, but a boat that might actually win will not show on the leader display as it might take many readings for their average speed to rise and project what might happen.* If readings are averaged over too few tracker data points the results bounce around a lot.* We choose a 12 hour average, or three readings.** Suppose boat X and Y are matched in rating. Suppose X is behind Y . The latest tracker update occurs for boat X 3 hours ahead of the Y, and X is suddenly blasting along at 10 knots, while Y is still mired in a hole after being run over by a squall.** Suddenly X's projected finish time looks golden, and Y's is still sitting at a relatively lower speed.* Now X pulls ahead, even though physically behind.* Added to this is that the tracker speed data in momentary.* Each reading, every four hours, is the instantaneous GPS reading, which is itself dubious for a single reading. That is a tracker limitation.
In the early part of the race the CPET was in 1000's of* hours for the boats stuck near the Farallon's.* At other points I have seen momentary speed readings that have pushed vessels way into the lead.
I hope that helps.* The race of course isn't scored until all the boats have crossed the finish line. The map projection is just a projection. Someday we could all discuss the merits of a variety of scoring algorithm's.** Something that looked out and judged the weather and had a factor for a skippers performance comes to mind.** The projections for finish time we are using at HQ are based on a gut call for each boats speed.* In two days we should see how this all worked out.
Regards,
Brian Boschma
The answerHEre is the issues:
1.* There is noise in the speed data.
2.* In essence timing is synchronized because the time to finish is based on the distance to the finish and then that time is added to the time on course.* So if two readings are 4 hours apart for X and Y, and Y was the latest entry, her distance to the island would be closer, that would be offset by her additional 4 hours on the course when the correction factor for rating is tossed into the calculation.** If we wanted this to be very non noisy we should have averaged more readings, but then you would still be wrong.* Boats that go stuck in the hole, for example, would look even worse than they do.
The prediction algorithm is fun because it is like discussing who is the best pitcher ever in baseball.Hello James,
The leader board compares the Corrected Projected Elapsed Time (CPET) of all boats in the race.*
CPET =* ((DTF/ProjectedSpeed) + TimeOnCourse) -* Rating*CourseDistance***** all adjusted to hours in this case.
The ProjectedSpeed is where the variability occurs. We have to select an algorithm to guess speed. If you average many readings of speed the results are less noisy, but a boat that might actually win will not show on the leader display as it might take many readings for their average speed to rise and project what might happen.* If readings are averaged over too few tracker data points the results bounce around a lot.* We choose a 12 hour average, or three readings.** Suppose boat X and Y are matched in rating. Suppose X is behind Y . The latest tracker update occurs for boat X 3 hours ahead of the Y, and X is suddenly blasting along at 10 knots, while Y is still mired in a hole after being run over by a squall.** Suddenly X's projected finish time looks golden, and Y's is still sitting at a relatively lower speed.* Now X pulls ahead, even though physically behind.* Added to this is that the tracker speed data in momentary.* Each reading, every four hours, is the instantaneous GPS reading, which is itself dubious for a single reading. That is a tracker limitation.
In the early part of the race the CPET was in 1000's of* hours for the boats stuck near the Farallon's.* At other points I have seen momentary speed readings that have pushed vessels way into the lead.
I hope that helps.* The race of course isn't scored until all the boats have crossed the finish line. The map projection is just a projection. Someday we could all discuss the merits of a variety of scoring algorithm's.** Something that looked out and judged the weather and had a factor for a skippers performance comes to mind.** The projections for finish time we are using at HQ are based on a gut call for each boats speed.* In two days we should see how this all worked out.
Regards,
Brian Boschma
Last edited:
hodgmo
Still Floating
Here's a synopsis from Pinnacle, the company managing the tracker page:
"the device updates are independent of each other, so one boat could report at 8:45 AM and the other boat could report at 10 AM. If a data report happens to be when one boat has a strong wind, that could weight the logic too much in one direction or another. These calculations are based on a single data point every 4 hours, so take it with a grain of salt. The horse-race aspect to the leaderboard is just some data processing we've added to the map. It's a formula based on the data received, and so I wouldn't consider any calculation as anything more than a rough approximation based on the limited and raw data."
As Brian says above, "The race of course isn't scored until all the boats have crossed the finish line."
"the device updates are independent of each other, so one boat could report at 8:45 AM and the other boat could report at 10 AM. If a data report happens to be when one boat has a strong wind, that could weight the logic too much in one direction or another. These calculations are based on a single data point every 4 hours, so take it with a grain of salt. The horse-race aspect to the leaderboard is just some data processing we've added to the map. It's a formula based on the data received, and so I wouldn't consider any calculation as anything more than a rough approximation based on the limited and raw data."
As Brian says above, "The race of course isn't scored until all the boats have crossed the finish line."
solosailor
New member
I understand there are glitches but to continuously show a faster boat ahead in the standings who is 30 miles behind a slower boat is quite an error, not just a few hours time variance, like 24 hours. I'm not sure how the other race trackers do it but I think it has to do with an "averaged" data set, not a single data point. Maybe a calc. if the race ended right now where would they stack up, that's a way better indicator and takes skippers and boats performance thus far, not a calc based on a few seconds of high speed surfing or wallowing between swells. If it's that off it should simply be left out.... maybe a leaderboard with only DTF, not corrected.
Last edited:
Jonathan Gutoff
New member
Ray has set up a nice tracker with estimates also. Same data in though...
https://www.jibeset.net/gpswatch.php
https://www.jibeset.net/gpswatch.php
brianb
Olson 34 Driver
Hi Greg,
If you want you can take the position reports and construct your own algorithm, or just manually make a projection, build and alternate leaderboard, and I will post for you on the SHTP website.
Regards,
Brian
If you want you can take the position reports and construct your own algorithm, or just manually make a projection, build and alternate leaderboard, and I will post for you on the SHTP website.
Regards,
Brian
solosailor
New member
As the public can see, but may not be able to understand, is the boat listed in 1st OA (Ventus) on the "Leaderboard" is physically 40nm behind behind a slower rated boat (Kato) who is listed 4th OA but really WAY ahead corrected. My guesstimate-o-meter says corrected is: Kato is ahead of Ventus. I'd need more coffee in my algorithm to do more.
Last edited:
Philpott
Cal 2-27 Dura Mater
The Kahuna chuckles. Plenty of coffee ingested at Command Central. Wish you were here.