Facebook

Login

Support Sailonline

If you haven't already - join the SAILONLINE YACHT CLUB!

Please also consider making a donation - all amounts are greatly appreciated!

Board » General Discussion » DTF calculation

Posted by
NZL_Scotsman


Users avatar

Posts
251
Registered at
2008-10-04 07:59:40
2012-05-01 09:57:51 - DTF calculation
One of the most frequently asked questions on chat is how the distance to finish is calculated. I've tried to explain that, based on my understanding. If anyone knows differently, please add to the thread.

The first two plots here are sketches of a race similar to our SWR Leg 6. That is a good example of where the DTF is not entirely intuitive. The third plot is the mark to mark distance, which many think should be the DTF.

The distance to the next mark is determined by the closest distance to the line that goes out from the mark. This line (the mark 2 rounding line) is drawn at right angles to the line between marks 1 and 3, starting at mark 2 and continuing infinitely.

From that point, the next part of the DTF is the distance to the next mark. Then mark to mark to the finish.

So looking at the attached plots. Boat A sails close to the rhumbline, so its DTF is very similar to a mark to mark course.

Boat B, however, is sailing wide of the mark. From B1 we find the closest point on the rounding line (point X). From there to mark 3.

All the distances from the various positions are shown on the plots.

--- Last Edited by NZL Scotsman at 2012-05-01 10:11:12 ---
Attachments
Posted by
NZL_Scotsman


Users avatar

Posts
251
Registered at
2008-10-04 07:59:40
2012-05-01 10:15:28 - Re: DTF calculation
If you have another look at the mark to mark system, the main problem with this is that when you approach the mark rounding line wide - then the DTF includes the distance to this mark until you pass it. Then it jumps to the next mark.

For example Boat B at B4 jumps from 43nm to 35nm as soon as the line is crossed.

The existing system has a similar problem, but it is much less severe. The problem only really occurs when you have a situation like B3, where the mark 3 rounding is very wide and therefore the mark 3 to 4 distance is much greater than compared to boat A. (35nm compared to 17nm).
Posted by
NZL_Scotsman


Users avatar

Posts
251
Registered at
2008-10-04 07:59:40
2012-05-01 10:25:24 - Re: DTF calculation
Here is an alternative method of calculating DTF that might avoid some of the issues the current version creates - those rare occasions when the course shape results in longer DTF than is reasonable.

The attached plot will help with this description. Instead of waiting until the mark is rounded before the next mark is considered, DTF is determined by going to the next mark as soon as the extended bearing is crossed. If the boat crosses back, then simply revert to distance to the mark.

For example B1 has crossed the line, therefore DTF is distance to Mark 3 plus 3-4 distance (141+17=158nm). This compares to 166nm with current method.

Now compare B2 and B3. The boat has only moved about 15nm. DTF for B2 = 108nm, B3 = 98nm. Sounds reasonable?

The existing method B2 - 112, B3 - 111.

I think this method will be easy to compute, so it will not add to the server requirements. I cannot think of any disadvantages, but might be worth testing it with some more examples.
Attachments
Posted by
AGage


Users avatar

Posts
411
Registered at
2008-12-27 10:31:57
2012-05-02 06:36:49 - Re: DTF calculation
My personal solution is to ignore ranking and sail, all will be sweet when we cross the line ;-)

Your method of calculating is slightly off. Example 1 shows best (at our current situation on VOR6) is how it is currently done, but rather than orthogonal to the line *1-*3, the 'scoring line' that extends from *2 to score roundings bisects the angle *1-*2-*3. OK, in your diagrams this does not make a lot of difference - but shift *2 along the direction of line *1-*3 and you can see it can... This is the BRG value on the NOR - Brg from ship to mark at the instant the server scores your rounding.

Just some advice from experience (regarding alternative methods) - you can draw 100 examples and it all works fine, but there is often some edge case where things go nuts...
Diagram 4 does appear a better method of computing DTF, but not of scoring roundings (B3 should not be considered to be on Leg3 for example, and could flip from Leg2 to 3 many times on its way to the finish? sim. B1) and we still need the line we currently have extending from the mark to score roundings (and to detect 'unwindings')
This would be where is creates more load to compute - 2 lines crossings per mark instead of 1 - actually 1 line crossing (score) and 1 'which side of the line are we' (dtf)
(not saying is a bad idea, just commenting because you did about server loads)

It's great that others are thinking on things like this though - when you are doing such examples though, try pulling out GE and drawing LONG courses, as these are really the only ones where the ranking ever becomes an issue for some - when a GC path significantly changes where on the line the scoring point is, compared to a Loxo (and also the scoring line itself, being GC) - small distances like in your diagrams don't show this effect. Also GC can change the edge cases a bit - remembering back to the ranking fiasco on Leg 5 while the scoring point was the south pole for pretty much every boat.

Of course, the other question is "Is this a good allocation of development resources?" - which always needs to be asked ;-)
I help develop the client interface for the best online ocean racing sim there is... __/)/)_/)__
Posted by
Amigo_do_Boteco1


Users avatar

Posts
4
Registered at
2008-01-18 11:07:00
2012-05-04 01:14:52 - Re: DTF calculation
Thank you NZL Scotsman

This is a very complicate way to measure the DTF.

I think its not that hard to implement a way to calculate the DTF based on a point to point measure:
- point to point for "mark to mark", and
- point to point for "boat to next mark".

Its a suggestion to SOL programers.

Fair winds

--- Last Edited by Armando Faria at 2012-05-04 02:33:41 ---
Boteco1 - Make Bore Impossible
Posted by
AGage


Users avatar

Posts
411
Registered at
2008-12-27 10:31:57
2012-05-11 06:49:06 - Re: DTF calculation
It is point to point for future legs??
It is actually point to point for all legs, just for the first 2 the exact location of that point is not fixed...

It isnt really any more complicated than just computing boat to mark to mark etc, but it is more realistic. It is exceedingly difficult to write an algorithm that can most accurately handle all possible course configurations, sizes, and locations (for example, had the VOR Leg5 been translated N so that it crossed the Equator, but otherwise remained the same shape, we would not have had the ranking issues with the South Pole)

If we change it then boats that stay close to the rhumbline will most likely figure higher in the rankings than they are. Personally I would rather appear lower in ranking than I really am. There is no realistic way to have a consistently accurate ranking across the whole race - given the very different configs/sizes/locations of SOL races and the variety of possible routes.

Rank calculations and visualisations like you see on the AC45s and VOR inport races only work for pure Windward/Leeward courses, and even then as the wind becomes more shifty they become less accurate.

I don't really see there is much wrong with the current system
I help develop the client interface for the best online ocean racing sim there is... __/)/)_/)__

Please login to post a reply.

Races

Next Race: 00d 00h 00m

Current Races:

New Ice Age 2026 - Browerville to Longyearbyen


Our 2026 long-distance ice-boating challenge continues with a second race in Arctic waters, now from Browerville, where we arrived less than a month ago, continuing east, back to Svalbard’s Longyearbyen, leaving Canada’s Ellesmere Island to port and Denmark’s Greenland to starboard. It’s about 2500nm, which, if there’s wind at all, should, well-wrapped in your DN-cockpits, not take you more than a day or three, thus to complete your first (of several) online circumnavigation of the planet of 2026!
Race #1969
INFOby brainaid.de
DN PARTICULARS
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking: NIA - SYC
Race starts: Feb 19th 07:00 Registration Open!
▶ Flash
GO TO RACE

Doha Basra TIMED Race 2026

Welcome to Kuwait for a short 78 nm TIMED race from the port of Doha (not to be confused with Doha, Qatar) to Basra, Iraq. Jazirat Light on Auha Island is the only mark to observe, leaving it to port. The Beneteau First 40.7v2 racer-cruiser is our ride this month. This is a TIMEDrace so you may RE-REGISTER HEREto try again after finishing a run. You will have 13 days and 11 hours to test your skill and decision making after the race opens.
Race #2008
INFOby brainaid.de
First 40.7v2 Particulars
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking:
TRQ1 - TRCH - SYC
RACE CLOSE: Saturday,
28 February at 23:00 UTC
Race starts: Feb 15th 12:00 Registration Open!
▶ Flash
GO TO RACE

Cross the Convergence 2026 - Samoa to Hawaii


The second race of our new series of ocean voyages across the World’s convergence zones, takes us northward again up the Pacific Ocean, now from Samoa at 13.5 degrees South, across the Equator to Hawaii, famed for its breaking surf and active volcanoes, on the edge of the Tropic of Cancer at 19.5 degrees North. It’s 2300nm, so we’ll take our very steady Steinlager II. Race #1982
INFOby brainaid.de
Steinlager IIPARTICULARS
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking: CCZ - SYC
RACE CLOSE: Saturday,
February 21 at 2300 UTC.
Race starts: Feb 06th 18:00 Registration Closed
▶ Flash
GO TO RACE

Beketov by Balloon 2026 - In Siberia


Way-back-when France and England, and Spain as well, were squabbling about who should control what of North America, only to lose the most of it to their own insurging colonists, their eastern European neighbour empire, Russia, was quietly assimilating vast, thinly nomadically populated territory of its own – Siberia! Amongst the many explorers and adventurers that served Russia so well in these conquests was the Cossack hetman Pyotr Beketov. Let’s pretend we’re he and eschewing horseback, let’s travel by balloon from Yekaterinberg to Krasnoyarsk – just a short 2000km (1100nm) section of the Trans-Siberian Express’ 9289km!
Race #1972
INFOby brainaid.de
SOL Balloon PARTICULARS
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking: BOL - SYC
RACE CLOSE: Wednesday,
March 4 at 2300 UTC.
Race starts: Feb 03rd 13:00 Registration Closed
▶ Flash
GO TO RACE

West Papua to Oregon 2026

Our next race in our Ocean Championships is the first of a series of great natural migrations we will track in 2026. The leatherback turtle is the most widely distributed marine reptile on planet Earth, and chooses to breed in warm tropical waters, but prefers to forage in more temperate habitats, travelling thousands and thousands of miles effortlessly annually to maintain this way of life. Our race will follow one of the typical trips of this turtle, from breeding grounds in the seas off West Papua to the coast of Oregon. It’s only 5900nm, so to keep up, but in comfort, we’ll follow in our Gunboat 90.
Race #2006
INFO by brainaid.de
GB 90 PARTICULARS
WX Updates:
0430 / 1030 / 1630 / 2230
Ranking: OCQ1 - OCCH - MIG - SYC
RACE CLOSE: Tuesday,
February 24 at 2300 UTC.
Race starts: Feb 02nd 11:00 Registration Closed
▶ Flash
GO TO RACE

Go to race archive

SYC Ranking

  1. Sailonline Yacht Club Member KaSToR
  2. Sailonline Yacht Club Member WRmirekd
  3. Sailonline Yacht Club Member CriticalHippo
  4. Sailonline Yacht Club Member rumskib
  5. Sailonline Yacht Club Member rafa
  6. Sailonline Yacht Club Member Sax747
  7. Sailonline Yacht Club Member vida
  8. Sailonline Yacht Club Member CollegeFund
  9. Sailonline Yacht Club Member bonknhoot
  10. Sailonline Yacht Club Member BRENTGRAY

View full list

Series

Mobile Client

SYC members have the benefit of access to our mobile/lightweight web client!

The mobile client