I just came back from a regatta where we used Racesense. It's a great tool. There are less recalls because you have the distance to line info on your screen. There's less guessing on where the line is especially if you are burried midline. We wound up ocs twice and restarted promptly. If the RO observes that the race is going pear shaped, he/she can always pull the plug and postpone. Not sure, but I believe it's the same thing if theres a general recall. There's a provision in the SIs that if your unit is not functional it's your fault and cannot be grounds for redress.

I have used both Racesense from Vakaros and RTK from Velocitek.  I like them both.  Both products have an option to allow the PRO to have a final say on OCSs before reporting them to the competitors.  I highly recommend that option.  I explained the intentions of the RC to all during the competitors' meetings.  As far as SIs, I simply added SI 9.4 https://ys-uploads-prod2.s3.us-west-1.amazonaws.com/event_documents/16424/ea584f82-bb21-436e-8892-0b098a72fa29.pdf for the Vakaros event.  I added a similar statement SI 9.5 https://ys-uploads-prod2.s3.us-west-1.amazonaws.com/event_documents/16419/ef5d7599-25ba-41bd-bf6b-3d1f706bd7c6.pdf for the Velocitek event.

I have used Vakaros Racesence for the 2024 WASZP North Americans and it was very successful. The RO based on the class rules can configure the amount of information to display on the boat's units.  It won't take them long for the boats to realize that they can't hide beside other boats to avoid being called OCS/BFD/UFD You'll find that boats will stay back of the line.  It should be noted that RRS 26 is changed because the start will be based on the device time and not the signals.  Clearly this is a very good direction for our sport.  The system I used sampled 10 times per second and this is what you need for high speed boats.  A 30 knots, a kite will sail 15 metres per second which shows that you need a 10Hz or greater sample rate.

I am so sad about this technology. This is my ONLY ELITE SKILL in sailboat racing! If everyone can start easily, I will go from mid-fleet to slightly worse than mid-fleet. I jest, but only as a coping method :-P

Speaking as an Internet software guy that monitors website outages, I can say that any product like Racesense a) depends on a wide range of back end internet services to work; and b) those services often experience outages. If any of those services don’t work, it’s likely the entire system doesn’t work.

For this reason, the NOR or SIs need to have some language so that there’s a fallback when Racesense isn’t functioning properly (or is completely down). Maybe that’s a signal flag stating that Racesense isn’t being used and the race committee is reverting to traditional techniques of manually sighting OCS boats, code flags for individual and general recalls, etc. 

No it does not. Racesense does not rely on 4G / Internet connection.
You could run racesense in the middle of the atlantic

From their website:
How does RaceSense communicate? Does it use cellular networks?

RaceSense utilizes a highly reliable, fault-tolerant mesh network for all race management communication on the water. This ensures RaceSense can provide accurate and reliable race management anywhere in the world. 

RaceSense specifically avoids use of cellular and other 3rd party networks for critical race management functions. On-the-water, these networks frequently have coverage gaps, high latency, and other challenges. We view RaceSense as a mission critical system, and consider these to be unacceptable risks.

I won’t be using it until someone can guarantee that it is actuate to inches, I will not take the chance of calling one back because of an error of inches in the system. Just is not fair. To many things that can cause an error and anyone who takes such gps instruments as totally good really doesn’t understand the system. Electronics are much better these days and pretty dependable but that kind of accuracy just is not dependable  

@Dennis With all due respect, current system is accurate within 25 cm (IOW inches). This is IMHO much more accurate that any RO bobbing on a signal vessel. Their new RTK system will be accurate within 1cm. As far as the OCS call is concerned, I have far more confidence in Racesense than a RO. I served as RO. I can guarantee that I was not accurate within 25 cm.

These kind of systems are more fair then officials looking over the start line at a start of 60 J70's. The uniform/black flag is a very harsh penalty for a competitor who is inches OCS. These kind of system allow start under P even with >60 boats. At this point I only can say listen to the sailors, they are not enjoying general recalls at all. These systems potential ends uniform and black flag starts, which we both do not enjoy as a sailor.

I have now done several events with Racesense and truly impressive.

https://www.instagram.com/reel/DAWKEDOtlSY/?igsh=aDdyaGpzZG1jM3Vx

Check this video out, Racesense in action. 
-Starting under P
-At the pin end a GBR-RS21
-GBR sees on his Vakaros "OCS" /w flashing red lights. Immediate at start signal.
-GBR has enough room to pin to correct and bears off immediate
-When GBR dips under line and is clear he will see 'CLEAR' on his Vakaros.
-Immediate turns back and is racing, second line start but still racing!!

If you see this, I mean.... what can I say as a sailor this i s how I would want it.
Myself sailing Star NED-8313, I certainly hope we have Racesense at Worlds in Kroatie this year (100 entries)

I've been thinking for months about how regatta's will change when (and this moment is not to far) this technology will be used for large regatta's. In Italy we have often regatta with 80/150 optimist or ilca on the same starting line, and in some events with young sailors is not impossible to have a fair start line with all the fleet out and 2/3/4 general recalls. How will change our work and how will be the transition for sailors and coaches? And then maybe a bigger problem: we need to completely change our methods of communicate the countdown with the sailors, we will use instruments with 1/10 of second of precision but the sailors will continue to follow our flags signals? I'm not against the process but I am curious on how sailing competitions will change

@Jacopo:
For years many fleets who allow full use of electronic tools have used start line identification devices from companies like Velocitek and various other GPS based tools to judge their proximity to a start line.
This doesn't really change any of that.
Also I don't really see this sort of tech being installed on an Optimist any time soon, although I could be wrong.

As for "endless recalls" there is a cure for that:  escalating to "Uniform", then Black. Or Indigo and/or Zulu, depending on the degree of control the RC deems appropriate. 30.1,.2,.3,.4.

Wouter, thanks for pointing out the mesh network. I'm sure that's very reliable, but it seems highly likely they don't operate completely offline, i.e. without any Internet connectivity.

The reason I say this is that, if you look at Vakaros' pricing model for RaceSense, a sailor buys an annual pass or monthly pass (https://vakaros.com/products/month-pass). That means you need to authenticate someone (e.g., Auth0 or similar service) and processes payments (e.g., Stripe or similar service). 

Let's suppose they use Auth0. Auth0 has had issues as recently as two weeks ago:
https://status.auth0.com/incidents

And let's suppose they use Stripe. Stripe's had issues the past couple of days:
https://status.stripe.com/

I'm just picking Auth0 and Stripe as representative vendors; any cloud vendor will have downtime. The point is to be prepared. So, race committees should have a fallback, such as traditional race signals.

I have absolutely no conflict here - nor any preference to a system.

I have been considering many of the above points recently, and was directed to this presentation which is useful regardless of which system being discussed.

Actually, many pertinent questions answered.  Thanks to the presenters.

https://youtu.be/EPSAhtVV70M?si=WWmYWMAnI8sXq3_B&sfnsn=wa

Hope it helps.

I wonder with the 25cm of accuracy and where the unit is placed on the boat how the OCS-border is determined. I have race with a number of software systems and apps on different boats and you always have to enter something als the length of your boat .... relative to the GPS placement. I did not see any instruction yet or SI phrase that states explicitely where the unit must be placed mandatory! Everyone may assume it is on the mast.... but is it?
Or can I have two units..... one visible on the mast (and not registered.....) and the othe on the back of my boat that is registered and not clearly visible for outsiders? Or can I "tinker" with my system and demount the GPS from inside and run a it with a wire to the back of my boat and mount the little receiver there? 

Just some thoughts .......   I prefer classic starting and just have the RC fly one or two drones static positioned over the staring line.  Plus I believe that system and software-license are too much of a burden for many sailors

If it is that accurate then are you using it for the finishes also.

@Gijs The shape/dimensions of the boat are programmed into the software. No need for other units or extensions thereof. When you set up the unit, you enter the offset distance from the bow.

To comment on one sentence in the original post: "Is a race sailed where a large fraction of the fleet is recalled really a fair race and a desirable outcome?"

Like many things: It depends.

If the PRO realizes that the large number of boats OCS is due to a poorly set line or radical changes in wind or current, then I feel that AP prior to the start or a General Recall is the correct response. 

However, if the large number of boats OCS is simply a fleet characteristic or early event jitters, then I do not agree. My reason is that within a fleet with a large number of OCS boats there are boats who did not break a rule and started on time. Every time I have to use a General Recall I have to accept that I've deprived the boat(s) that obeyed the rules of their right to win the start. That seems fundamentally unfair to those who obeyed the rules and is to be avoided.

@ Alain "When you set up the unit, you enter the offset distance from the bow" 
That is exactly my point! Sailors, similar to F1 teams, have the habit of "tinkering" with class rules and Ratings all the time. Many search the grey-limits.....  So every sailor has to manually enter the offset dinstance from the bow. For me that would immediately tempt me to make a negative correction of at least 25cm, since that is the accuracy tolerance mentioned. If there are waves, I might even add a few centimeters more..... And the real neirds will open up the Vakaros and definitely look if they can find a way to dismount the GPS-chip and rewire it so it can be placed aft of the unit itself, which ofcourse has the offsett nicely entered for the position of the unit .....  No one will notice, since llikely the whole RC is only looking down at a screen..... 

I am just letting my imagination go free here ..... Trust me, I have seen plenty of creative ways to tinker with class and rating regulations which for some is a sport in itself. 

@ Beau - I fully agree with your input!

Thanks for the input, looks like a few people have answered other's questions too.  

Like most threads, this one has interesting drift. (By way of background, I'm a retired tech guy and investor with a long history in the semiconductor and software industries. I'm also a retired Regional Race Officer and National Judge.) There are some real issues that a Judge and Race Officer must consider when thinking about these systems. More on that at the end.

The Future: With respect to what can happen in the future, all these systems should eventually:

- Notify when a boat is in the Zone around a mark
- Notify the Competitors when there is an overlap, and when it is broken
- Log the finish time of each boat and time of each mark rounding for use in applying redress and safety monitoring
- Notify the Competitors and Log any violation of restricted areas called out in the NoR and SIs
- Provide real-time location tracking without the need for other trackers (depending on $$$s)
- Provide a file to the Jury of the position of all Boats, Marks, Boat Speed, Course Tracks, Headings, etc... during the race and for analysis by the competitors.
- Add something I haven't thought of yet.....

For a good look at what the future will be for all racers, one need only review what Stan Honey and his team have done for the America's Cup and SailGP. Eventually this is foundation the same sort of system fulfilling Stan's forecast that eventually you can do all this with a phone. If one adds wind speed and direction sensors, even the grids lines that appeared for the AC and SailGP can be superimposed on the data files for races for real-time viewing and post-race review. We are not at the "...with a phone." state yet because of the terrible antenna in the phone and in most consumer grade GPS devices.

Accuracy and Quality of an Antenna: All Race Officers, Jury Members, and Competitors need to educate themselves on the true real-world accuracy of these systems. Sadly, some folks are conflating the specified accuracy of the GPS chips with the actual accuracy of the system in which the chip is embedded. The "chip" used for RTK-level GPS can be accurate down to 2cm (0.79"). However, to achieve this accuracy, one must provide a significantly larger antenna to the RTK-level GPS Chip than any current vendor provides outside of expensive and quite large separate devices. It appears that the Minimum size antenna required is currently around 10cm (4") in diameter, it must have a comparatively unobstructed view of the sky to receive a minimum number of satellites, and it currently takes substantially more power than cell phone vendors and other devices are willing to "pay" (thus a bigger battery). What this means is that a new generation of high accuracy RTK-GPS devices (meaning displays with GPS in them) must be developed before anyone can reach the 2cm accuracy level that the highest competitors could require. For some of the current tests of RTK-GPS enabled starting line systems in places like the New York and St. Francis yacht clubs, large external commercial RTK-GPS devices have been used to achieve this accuracy. ROs and Judges need to understand what vendors are talking about out when they make a claim of accuracy.

Measuring Accuracy: Like many things in the real world there is a distribution of results associated with accuracy. The current generation of integrated phone style GPS devices has a tested accuracy distribution of 1 meters for a given level of repeatability. Like most mathematical distributions, the maximum error rate is far less that 1 meter, but Judges and ROs need to know that in a large fleet (100s of boats) across multiple races at an event, there is a significant possibility of at least 0.5 to 1 meter error occurring. While this is still much better than the accuracy of a bow-crew bouncing up and down on the pulpit trying to call the line, it is not more accurate than a PRO's team sighting the line. 

Like many things in the real world, the probability of errors is distributed and represented by a curve. Therefore, a completely accurate measure of a device's "accuracy" would be expressed by a probability of some unacceptable error expressed as something like "1 in 1,000 for a 10cm error". A simple statement of accuracy without probability can be misleading. For the America's Cup and SailGP, an accuracy of 2cm was not only required, it was regularly tested by competitors. Put another way, those competitors were hitting the line with speed and often within single digit centimeters. I believe one would find that calibre of accuracy demanded by Olympic level fleets and many other Classes that are highly competitive. (EG: Star, ILCA, Moth, Etchells) For these reasons, some of us have been advising OAs and PROs on these issues and clarifying expectations regarding current solutions true accuracy when run without a RTK-GPS with a large antenna.

Hi, I used RaceSense in a 40 boat sportsboat fleet Worlds last year. 30 professional boats. Autonomous marks. I also had a team calling the start as usual and also calling the finishes, which were also videoed. 

On the first race I had 3 boats "go dark" a couple of minutes before the start and appear after the start. The system told me who they were. I said I would abandon any race where this happened again and give them new units. It never happened again. 

I was told that each unit was preprogrammed with the shape of the boat. I don't know if the competitors could change that nor if it was programmed as a box or a boat shape. 
I was also told that at a previous regatta a couple of boats had been wrongly programmed as TP52s so were OCS each race on that day. 

The Race Docs allowed the RC & PC to request redress but not the competitor. We had a flag we could put up to go back to normal starts if the system failed. It didn't. 

We had 9 races, all on P flag, 2 restarts, one because the Committee boat dragged  in the last 30 seconds putting a third of the fleet over in 20+ knots, the other on the penultimate start when the tide was under the fleet and most boats were over and I was concerned that the returning boats wold disadvantage the ones not over and there was a serious risk of accidents. The re-start was clean. 

All the boats my start team identified as over were recalled, however there was often 3-6 boats called OCS that my start team thought had started properly. The system reported them all clear within a second. 

The finishing order differed substantially both from my finish team results and the video. Boats coming up from behind going fast were often called as finished in front of slower boats in front. 

From a sailor's point of view most boats ended up hovering on the line at the start. The boats that did well either had room to bear away down the line to keep their speed up or hit the line going fast just after the start and popped out in front of of the others who were going slowly.

In summary I liked the system. We had no OCS, UFD or BFDs and only 2 re-starts. Sailors enjoyed the event. The best sailors still won, the game is just slightly different. Personally I wouldn't bother with starting flags as un-necessary unless the system fails. The competitors need briefing properly. I would love to know more about who sets the boat "type" in the boat unit and who can change it. 

Happy to share my race document wording if anyone wants it. 

The discussion has been interesting. 
RRS 56.3 in this rule book includes new wording "or any other tracking device, it shall not be turned off to its effectiveness intentionally reduced." My question is: are the race officials willing to protest a boat who turns off their device? 

Personally, I would have no problem protesting the boat.  I would be very motivated to do so and I would be seriously looking at rules 2 & 69 quite closely.  There is no excuse for intentionally turning off a required tracking device.

I have had a lot more problems with this on offshore races than RaceSense. Always difficult to prove it wasn’t a technical problem with the unit, esp offshore. If AIS and the tracker both stop transmitting at the same time it is very suspicious. I always call a hearing and want to ask what happened as well as see their course on the chartplotter to make sure they haven’t entered any restricted areas. Starllink helps as you can see if they are online. 
With RaceSense l tell them in the briefing I will stop the start if they disappear and replace their unit. If it happens again I will score them DNS (or OCS if I saw them over). Never had to do that in practice. 

About the rule 56.3 as jonh said i woldn't have any problem to protest a boat when the raceSence is tourned off. Coming back to the first question (is a race fair with the 50%+ of the fleet is OCS?) maybe policies will consider this problem in future but with all the front line OCS, that need to bear away and return under the line, how they can't interfere the fairness the non OCS boats, so i will consider to AP the procedure or even abbandon the race after the start.

For most people, opening the Vakaros unit is impossible. I recently purchased one as part of a Circular Error Probability (CEP) comparison between traditional and autonomous marks. The unit is completely sealed and charges wirelessly.

A sailor who opens and modifies these devices is likely the same type of person who cheats in other ways, such as:

In such cases, I believe a RRS 69 report would be the appropriate course of action.

There was a discussion regarding accuracy, and I believe that since all units operate in the same geographic area, their accuracy should be the same. However, what truly matters is precision, and devices like Vakaros have a high level of precision.

Here is a link to a document explaining the difference between accuracy and precision:   https://github.com/olpossum/circular_error_probable/blob/master/apn029.pdf    It also describes Circular Error Probability (CEP) in more detail.

By definition a boat starts when, her hull having been entirely on the pre-start side of the starting line at or after her starting signal any part of her hull crosses the starting line from the pre-start side of the course. A Definition may not be changed except by WS.

I remain to be convinced that the use of a tracking device complies with this definition. For instance can the system judge a boat sailing sideways, of backwards, across the line in a strong tide I look forward to reading the new Test Rule!

Gordon, the system knows the exact location of the entire hull. So far, it doesn't have a way to know where the rig is located, but includes fixed bowsprits and bowsprits in "normal position" for sailing upwind. The system most definitely knows if a boat is sailing backwards or sideways. It has a compass that is highly accurate and can provide "heading" data, it also can provide the "Course Of Ground" to a high level of accuracy, so leeway is easy to calculate, as is sailing backwards.

Peter, it is not the case that all of these units have the same level of "accuracy", as I said above accuracy is determined by the combination of the GPS chip used (technically is it a L1-L5 chip or not) in combination with the size and quality of the antenna employed. While many of these units, including devices from B&G, Garmin, and Raytheon, utilize similar chips which are primarily produced for cell phones, they do not have the same antennas. As a result, they have wildly different levels of accuracy.

At this point only the Velocitek prototype used by NYYC and StFYC, which is using an external antenna that is accurate to 2cm, can achieve that level of accuracy. The older Velocitek and Vakaros systems have a much lower level of accuracy to about 1 meter.

You are correct to point out that the precision of these devices is not being discussed here. So far, I have had no input from the chip manufacturers or any of the vendor that there are problems with precision, only that the errors created by poor antennas has reduce accuracy. You are also correct that both factors contribute to the size of the error. But, given that almost all of the vendors are using the same chips (used in cell phones), it's unlikely that the chip vendor has the precision wrong. (That last bit is just my opinion.)

Accuracy of this has yet been to be explained how it arrives at its detailed positioning to the stated cm accuracy. One tried to explain that a logarithmic formal is calculated from one position to another position the generates speed to calculate position at time of start. Problem with that is the inherent error in positioning and the slowing down of a vessel to keep from going over or the increase of speed of one that is now trying to get to the line. If this is part of the system then there is a big question as to accuracy. Then is the sensor transmitting from the bow or midship. Would be great to have this outlined because it all affect the line call.

For those of you who might wish to take a deeper dive into how the higher accuracy of RTK corrected GPS can reach centimeter accuracy, here is a great link to a description of the technology.

https://globalgpssystems.com/gnss/rtk-gps-understanding-real-time-kinematic-gps-technology/

When I referred to "all units," I meant the identical units used for the race. It would be interesting to compare the precision and accuracy of my old GARMIN 12 from 30 years ago with my Vakaros.

All this technology for starting races became possible after the definition of "Start" changed in 2021. The software incorporates the boat’s hull geometry, as defined in Section D of the Equipment Rules of Sailing.

We’ve been discussing race starts, but due to the relatively low sample rate, I believe determining the finish position is much more challenging. They likely use some form of extrapolation based on coordinate samples taken just before the finish line.

Beau thank you for the link and in reading the whole link one does find as has been mentioned the possible intermittent problems that can occur because of the signals use for the system. This is what many have mentioned but have been trying to say doesn’t happen but yet here it brings out what can happen to make it not be so accurate. To say it is totally dependable is a false statement. Thus is a reason many have commented on not depending on it. This not to say it isn’t something not to be used but that it is not 100% a cute at all time and can give a false reading as to a position of a vessel.

Dennis -- I guess I don't understand exactly what "possible intermittent problems" you're referring to. Yes, if one uses a cell-phone as the connection between the reference receiver and the mobile receiver, there is a chance of lost or delayed communication. However, both the Velocitek and Vakaros systems avoid using a cell-phone for precisely this reason. They use a high-reliability private network without these problems.

Finally, the accuracy we're talking about was used in the last two America's Cup events and all the SailGP events to-date and continues to run without failure. If it's good enough for these events, I feel it's certainly good enough for most regattas. 

J24 Worlds in Seattle used RaceSense for its many starts. There was only one instance to my eye where it erred - in a boat’s favor.  
Though it was not actually used for finishes, it was consulted to fix a disagreement between top deck and wheelhouse line callers. 
Also quite a helper in finding boats that had not checked in. 

Beau, I am talking about what is written in the link you shared. It does point out as anyone in electronics knows line of sight transmission and reception which can be affected by several things and it takes several lines of different angles to get a position accurate. Now in big regattas this can be avoided by the regatta establishing their own base stations or being in a position to not have any blockages of signal which would make them very accurate. Not going to say they can’t be good but there is the possibility with out your own bases there could be a problem. Not saying there will be but the chances are there as the link talks about. Interesting concept though.

Should the use of this system be included in NoR? 
What wording have you used?

Hi,

The following is the links to the Documents the RS21 Class uses. 

The Series and Event NoR is: https://www.racingrulesofsailing.org/events/10765
The standard Sailing Instructions are here (note para 32): https://s3.amazonaws.com/rrs.prod/vkoq3fzabjwkjvloucg8vt0e8ge2?response-content-disposition=inline%3B%20filename%3D%22RS21UK%2526ICup2025-SSIs%20Final.pdf%22%3B%20filename%2A%3DUTF-8%27%27RS21UK%26ICup2025-SSIs%2520Final.pdf&response-content-type=application%2Fpdf&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAJH37WFGGNKRSBH6Q%2F20250317%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20250317T123054Z&X-Amz-Expires=604800&X-Amz-SignedHeaders=host&X-Amz-Signature=4b0562c8f7e02790a626dde82dfeadd30c61cf6da72f8517a066733f6333a602

Hopefully those links work OK. 

Roger, 
1. Is there a typo in 32.7?
32.7 If the RC vessel displays flag GOLF with a sound signal before the warning signal of a race, SIs 21.1 and 21.2 will not apply and the Orange Flag
shall be displayed before the Warning Signal for the next race.

SI 21 refers to Event Advertising.

2. Do the SIs replace the visual and sound signals in RRS 26 by the system signals?

The future is here. Facts from Vakaros:

The J/70ICA will run RaceSense at all major events, including Worlds, continental championships, etc, as well as almost the entire domestic slate.

The Melges 15 class requires Atlas use at all major events, and most minor events. RaceSense is actually written into their class rules.

We are also the 'official race management technology' of the M32 Class, 69F, RS21 Italian Class (30-40 boats every event), Etchells Winter Series, Bacardi Cup, 2024 J24 Worlds, SailGP Inspire, and more that I'm inevitably forgetting.

RaceSense has been used at 14 World Championships, in classes including Etchells (2x), J/70, 49er, FX, Nacra, and more. 

So far in 2025, 80% of RaceSense events are NOT run by Vakaros employees or with VK staff onsite. Clubs and classes self operate as the system has become simpler. 

Winter Series: RS used at the Winter Series for the following classes/events: Bacardi, J/70, Melges 24, Melges 15, MC Scow, E Scow, Viper, VX One, Monaco J/70 Winter Series, Etchells, Great Lakes 52, Star, Lightning

We anticipate 150-200 domestic RaceSense events in 2025, including the Thistle Nationals, J/70 Corinthian Worlds/NAs

Internationally, we are also slated to run World Championships this year in the J/70, Star, Melges 24, M32, RS21, RS500, 18 foot skiff, and more, and have already run the 2025 Etchells Worlds.

Denis, I now understand. Sorry I didn't get the gist of your earlier comment. Here's the deal.

If you don't have your "own" base station, you most certainly can have problems. At least at the current time. In the near future, there could be commercially available base stations because of the demand from self-driving cars. They wish to know exactly where they are and current GPS, even with a good antenna, has too large a chance of error putting the car in the wrong lane. Mounted on poles or buildings, the base stations can have a much clearer view of the sky and reduce dead spots and multi-path effects. RTK GPS has a chance of possibly eliminating having to use camera/lidar for lane following to the degree it is used now. The car will still need to watch for changes that aren't yet mapped. etc...

The RTK GPS used in sailboat racing (America's Cup and SailGP) has always had a dedicated base station within range of a proprietary network that communicates to the Signal Boat and the Competitors, to address this problem. It is my understanding that all the significant vendors are working to move to RTK. Both NYYC and St. Francis YC's tests of RTK GPS included this. Tests that were not RTK obviously didn't have a base station at all. So the choice is basically reasonable probability of 2cm accuracy with RTK and about 1meter without it.

Lots of interesting discussion here. Two CM accuracy is not going to be achieved by GPS alone as issues in the propagation medium (the air) will impact the signals. To get this accuracy a corrector signal must be transmitted from a transmitter who's location is know precisely. This is not going to happen for all but the highest end regattas. 

But we don't need that level of accuracy, almost always the 25cm claimed is good enough. As others have said, the shape of the boat is programmed in and the placement of the device is mandated so its the same on every boat. Even with 25cm accuracy I believe that on balance, this system is more fair than humans looking down a line:

1 - 25cm is probably better than a human can site a line in many conditions
2 - No boat can hide
3 - The Atlas 2 or Edge device counts down and signals go. There is no delay from the speed of sound going down a long start line. 
4 - The competitor who is OCS is told instantly. Without these, a competitor may be called 6th ot 1st or 8th and it's unfair when one boat is called earlier than another. Same with being called clear. 
5 - Its not going to misread a number

Will it make an occasional mistake? Maybe, but on balance it's going to make far fewer errors than humans. 

Can it be cheated? The RC knows if a competitor turns it off. This is logged. If a unit fails, the nature of the failure will be logged and an official can examine the log to see if it was deliberately turned off or if there was a failure. It would take some strong hacking skills to hack into it, and even then you'll be able to get minimal advantage. Also each Atlas 3 has a 1 km range and together they form a mesh network. 

In addition to the advantages above, this system makes it much easier to put a race on. Each end of a start line has an Atlas 2 on it, the same unit that would be on a boat. The PRO has an iPad. One guy can start races and finish them with better management than a squad of people. Heck, a PRO could be on a dock; he or she does not need to see down the line. Get several robot marks and this system and we could probably have one guy run a race without a signal boat. 

There are other reasons this system is amazing. The primary downside is cost. Each Atlas 2 is $1,200 and an Edge is $750. Each competitor has to have one or the other. Then there are license fees. And what happens when a competing system comes out? Now the cost for each competitor goes even higher if they need one device for each popular system. 

Anyhow, my Atlas 2 is on backorder and I can't wait to get it! I'll be collecting some great data from my practice sessions.



 

Jim,

I will tell you from direct conversations with the Race Office at NYYC and StFYC that 25cm accuracy is not adequate for their events. Also, it won't be acceptable once a 2cm system is readily available. These Race Officers have seen 25cm is a significant error while serving as PRO of race events.

As to propagation delays etc... the base station corrects for that. Yes, a base station is needed, but it isn't needed in really time. The correction information is sent to all the devices through a high reliability local mesh network. In many areas this information is provided from ground stations as a service. I do not believe it changes rapidly and doesn't need to be up to the second. The services exist for other purposes, EG: guidance of self driving automobiles and location finding by surveyors.

Just as 25cm isn't accurate enough for a surveyor placing the property line before building a significant building, it isn't accurate enough for a Race Committee running something like SailGP, American's, Cup and I personally believe classes that are highly competitive like the Star. But, the Classes and Organizing Authorities will make their own decisions.

As to someone else's earlier comment about multiple systems on a mark or boat, I strongly believe it is in the best interest of our sport and especially the competitors that the various vendors all conform to the same system. This is not an area were differentiation will be helpful to anyone in the long term, especially the boat owners. It seems obvious that the suppliers of technology are racing to be "good enough" or "best", but it will be absurd if B&G, Garmin, Vakaros, Velocitek, Raymarine, and a host of others all have slightly different systems. I would strongly suggest that the Organizing Authorities and the Fleets demand standardization as it seems obvious that all boats from Optis to 100' sloops will need to comply in the next few years. 

Personally and privately, I have already heard from a number of sources that they aren't going to spend money on anything other than tests until the vendors get their act together to standardize. This is not an area where there will be any first-mover advantage to a vendor longer term, that will end up resulting in the customer who adopt to early being stranded with outdated technology in only a year or two. The above is only my opinion, obviously. But after decades in the tech industry, and watching numerous early innovators ignore the damage done by non-standard tech, I feel strongly that the vendors need to join forces to do this right.

25 cm is 10 inches, almost one foot. That’s a significant amount in a dinghy, and visible to the human eye, especially when starting near one end of the line. 

Put another way: 10 inches is 11% the length of an Opti, which is 90 inches long. It would be a significant advantage for anyone able to get bow out on the boats around them without penalty, and conversely, unfair to someone who was called over when they were in fact hanging back. 

Al, I agree 10 inches is enough to cause serious concern. As I've said elsewhere there is a distribution curve of probabilities of error. Depending on how much risk one wishes to take the error of GPS without a decent antenna and RTK is between 25cm and One Meter. Different vendors make different claims based primarily on where they draw the line of probability of an error that is acceptable. Most sailors would say a hard "NO" to 1-meter errors, fewer say a hard "NO" to 25cm, and almost everyone I've talked to about this is happy with 2cm, which can be accomplished with a decent sized antenna and RTK. 

Note that almost all vendors believe that RTK is required for 2cm accuracy. Various sources have assured me that to reach 2cm accuracy, one will need an antenna that is about 10cm long. That obviously means that none of the antenna in cell phone chipsets will get one to 2cm accuracy.

This is a fascinating discussion.
Beau, does a coiled or zig-zag antenna count as 10cm long?

Niko -- I'm not a qualified radio antenna engineer. I am relying on information from those who are. I'm also observing that commercially available RTK receivers are at least 4" apx 10cm in size. I'm pretty sure they'd like to make them smaller if they could. Google RTK GPS receivers, there are a number of them available that work for surveying. (An activity that needs extremely high accuracy.)

I have no objection to making the potential error smaller, but as I think about a potential 10 inch error, which is probably closer to 5 inches on average, I find that an acceptable tradeoff as compared to what happens when humans are sighting the line. At the Florida Midwinter Melges 15 regattas, there were 90 boats on one line. 

I know y'all are race professionals and I don't dispute that you are all very good at what you do. But there is no way a human is going to catch all the OCS boats in a 90 boat fleet, and then announce them promptly. Some boats are going to get away with hiding, probably more than just a few. The OCS boats that are announced last have a disadvantage as compared to those announced first. 

Also, such a long line is going to have sound propagation delay. Now if the horn is broadcast over VHF, that's fine, but otherwise there is a few seconds of delay between the horn being sounded on the signal boat and the horn being heard at the pin. 

So given all this, I accept an average 5 inch error as the price of getting rid of these other issues. 

Someone mentioned talking with people at NYYC and St Francis YC. Guys, there is a LOT of money there. I'm sure they can afford the error correcting systems and get down to 2 cm. When it comes to America's Cup and racing at a very high level, by all means spend the money. But for the vast majority of racing, there is just not that much money available for very expensive systems. We're already asking dinghy sailors to spend $1,200 on an Atlas 2 (okay, the Edge is $750). In the grand scheme of my J/111, that $1,200 is not a major issue. In the grand scheme of my Melges 15, yeah it is. Extra money spent on race management will translate into higher entry fees. 

As for standardization, I don't see the vendors going for it. I once worked for the market leader of a particular tech product. Years ago, there was a movement to standardize how software talked to this product. A committee of all the manufacturers was formed, and the company I worked for sent a representative. His instructions were to delay the effort as much as possible while appearing to support it. I would like to see it, but I just don't think we will soon. 

The manager of three Olympic classes told me recently that 30 cm accuracy is sufficient for his classes. Consistency and reducing the delays between races are very important.

I've been a line judge at over a hundred World, Continental, and National events, and if you think a race official can call a 300–400 metre line with 25 cm accuracy, you are mistaken. Line judges do the best they can and try to make sound judgment calls.

Below are some actual situations we face when calling the line by the rules — though I doubt many of us truly adjust for them:

Having the vendors work together would be nice, but I doubt that it will happen.  At the most recent US Sailing Leadership Forum, I attended a panel discussion covering this technology. I was surprised to see that the only company with a commercially available and actively used solution was not part of the panel — even though a representative from the vendor was in the audience, and the company had a booth in the exhibit hall.

Peter,

I would point out that there is another viewpoint regarding accuracy. Consider two star boats that are side by side, one slightly ahead. The one ahead is not called OCS by the electronics and the one behind is called OCS. Both competitors immediately lose faith in the system. That is how the competitors will judge the accuracy required. I'm sure you can appreciate the requests for redress in this situation. If both competitors answer in support of the systems error, obviously a jury would award redress to correct it. But, the damage to the credibility of the electronic system will have been done.

I agree completely with your comments about the movement of line-flags and the ability of a line judge to call the line to an accuracy of 30cm. I don't think it addressed the above.

We are arguing about centimetres. At a recent event, in heavy weather, with a big swell the committee boat was moving backwards and forwards by 6-8 metres!

@Beau, has such an incident actually happened?

@Beau, @Jim, in my experience as a racer racing under Race Sense,  whenever we got called OCS it was clear that we were in front of boats around us. Whenever others around us got called back is was clear they were ahead of us. Reality is, in a dense fleet without an instrument giving us where the line is, we usually do not have a clear picture of where the line actually is. You just pull the trigger based on the timer and other boats around you. I have raced in large Star fleets with a 1km long starting line. You do take your transits and such but in the last 3-5 seconds, helming with the bow about 12-14 feet in front of you, it's pretty hard to tell where the line is. As Peter van Muyden quoted in his post «Consistency and reducing the delays between races are very important» As racers this is what we appreciate from Race Sense. No general recalls, swift transition from one race to thet other, more races per day.

A timely article https://www.sailingworld.com/racing/starting-systems-are-changing-the-game/?oly_enc_id=0906C1421378J0U

Mark Foster is the best race officer I know and has extensive experience with Varkaros, Atlas 2 and RaceSense.  He is a no nonsense race officer who is a solid believer in RaceSense. Feedback from the competitors is even more compelling, they love it. No general recalls, no more wasted time repinging after a wind shift, just racing. J70, M15, Lightning have adopted, and now the Thistle class is adopting the system. Atlas 2s are backordered.

I believe the sailors have spoken.

Yes the Atlas 2 is backordered, I have one on order and can't wait for it to arrive! I'm literally checking the latest status every two days. And it never changes! 

Racesense start RS21 Italian Class last week in Sicily. this is what the majority of competitors want. 

https://www.facebook.com/share/r/19aE29zWxP/?mibextid=wwXIfr

I'm just catching up on all this. There were two situations raised which illustrate why 25cm is not accurate enough.

Finish Line - In this case the RC can almost always see the distance between boats, as there are far fewer of them finishing at the same time and the finish line is often much shorter to allow for more human accuracy. In addition, for downwind finishes with planing/foiling boats, the boats can be crossing the line at a speed that makes a human visual determination quite difficult. For these reasons it's common to video tape the finish and if a protest or request for redress arises for that video to be reviewed as evidence by the Jury. All of this makes it clear that 25cm is not accurate enough for the highest level of racing.

Determining Overlap - Obviously, this technology will soon be applied to determine exactly when a boat entered the zone around a mark and if boats were overlapped at that time. The difficulty with 25cm accuracy is that the bowman or someone on the transom (potentially wearing a video camera to prove his point) may show up with video that makes it clear that two boats were not overlapped while the electronic system says they were. (Exactly this pattern showed up in a recent high-level match racing regatta where the on-line and real-time video showed that the umpires were out of position and got a call wrong.) While the competitors may forgive a system that is accurate to 2cm (0.787") in my opinion they will not forgive a system that is inaccurate by 25cm (9.8"). (This technology can also determine if a boat has crossed a restricted boundary defined in the SIs.)

A final comment about other places this technology can help: In a future hearing regarding a port/starboard crossing where Port has claimed that Starboard was altering course, the system will be able to prove that to be true or false with a level of accuracy that is compelling. This has already occurred in some very high-level protest hearings in premier events with boats that have massively expensive and accurate compasses. There's no reason to believe that it won't happen as soon as these RTK based systems are available as highly accurate speed, heading, and rate of change could be recorded.

My primary issue continues to be that we're letting a robot sail our boat for us. To be clear, if boats are allowed to ping the line with similar manual tech, the issue is already moot. 

While I have no issue with the RC having this technology, I do think it takes away VASTLY from the skill required to start well. I already notice this in fleets where Vakros and other pinging-type machines are used, vs when they're not.
A similar, but unequal example is the MLB strike-zone. Accurate strike zones enhance the skills of players who have a good eye. This would be true for sailors were it not for sailors themselves having access to that info in real time.

I'd suggest that the issue of electronic assistance is already moot in many cases. All the high-end electronics provide significant assistance to tacticians and skippers when starting. But, I believes that illustrates the correct path. Each Organizing Authority and Class should be allowed to chose whatever they want. There is no need for the RRS to opine on what is "legal" for sailboat racing. This is basically the current model and it seems to work well. We have Classes and OAs that don't allow anything other than a compass (if that) and those that have almost no limits on what may be used (EG: on-shore routing and weather forecasting). Each sailor gets to choose the sort of competition in which they wish to participate. 

We are seeing a great example of this in real time with Rule 41(c). Many OAs are basically waiving 41(c) for two reasons. First, the spectators want to watch the event and that means putting all he information on-line. Second, the reality of the situation is that many competitors have a phone in their pocket that continuously allows access to wind and current forecasts, along with the position of their competitors. Having personally explained to people that when racing our boat in certain regattas they are not allowed to look at various subscription service versions of on-line weather and current forecasts, I've been shocked at how few realizes that could be a violation of Rule 41(c). To recognize reality, OAs are simply turning off Rule 41(c). (Indeed the term "freely available" has been reinterpreted to mean "all sailors are free to pay for a service", not its intended meaning, IMHO)

If one accepts that the OA and Class Rules define what is acceptable, within the scope of the RSS's restrictions, then we race officers and judges should simply do what they have asked of us. This is recognizing that some of us may feel that what OAs and Classes are doing isn't "good for the sport", but that's a separate discussion.

Yes, I agree Beau. My only hope is that some classes keep it analog...
I suppose someone could bemoan the use of digital watches (or countdown watches at all) in the same way I'm doing with line sight data.

Tomorrow I'll start an offshore race. I'm navigator.

My sextant is packed along with my hourglass.  I must remember to wind my watch up and sync with the harbour noon day gun so I'm not late for tonight's skippers' briefing.

Wait a minute - it's 2025! I'm not worried about where I am on the globe or what time it is.   I have at least 16 sources of accurate location and time on the boat. 

The navigator's game has changed over time with technology. It won't go back.  Not much can be done to stop it. Not much point trying to stop it. 

Sure, some classes and races can stipulate that 'old methods' must be retained, but the reality is that technology changes the sport and with it changes the skill set needed to win. 

Now my challenge is not knowing where we are, but to better interpret the information available to all to beat my competitors.

Perhaps by thinning out some skills, the game is more focused on other skills. e.g. If everyone starts at the same time, it's more important to sail well to win. 

So for me technology changing the sport is simply inevitable.

I don't see it as robots sailing for us. More like, one area of competition (whether it be chance or skill) removed, shifting focus on honing skills in others. 

Before weighing in on the pros or cons of these systems, I'd like to suggest that people familiarize themselves with what they can actually do.

It turns out that both Vakaros and Velocitek can turn off distance to the starting line, and show when a boat is OCS at the start. Source:

https://www.perplexity.ai/search/when-using-vakaros-instruments-lMalERibSReD76K4YKYcTw
(Perplexity links to its source documents if you want to drill in further.)

For those who believe that an essential part of our sport is accurately judging distance to the line, race organizers can turn off the distance to the line display. Because all OCS boats are called (albeit with +/- 25 cm or 1 foot accuracy) there are fewer general recalls, and less use of "windward marshmallows" (slow boats to windward) to hide from the race committee. 

To me, this seems like the best of both worlds: emphasizing traditional racing skills while promoting fairness. What do you think?

It's for sure not robots sailing the boats, although there some classes and races where an auto pilot is allowed. But even that is not a robot sailing the boat. In the case of the Melges 15 class, we're allowed to use DTL but not TTL. So I get a little help starting, but the Atlas 2 won't be driving through waves, locating the breeze, plotting the best strategy to the next mark, making tactical decisions when I meet other boats, doing any of the fundamental boat handling tasks. 

The winter M15 series in FL has more than 90 boats on the line at the same time. The guys up at NYYC can be all traditional because they have what, maybe 15 IC37s on a line? Humans can handle 15 boats on a line in most conditions with reasonable accuracy. No way they can handle 90. It's too easy to hide, and it would take forever for the RC to recall a bunch of OCS boats anyhow. Not every tool is right for every job. I see this tool as right for big OD fleets, especially performance boats. 

Al,

My thought is that the OA and the Class specify what they want the technology to do and what's legal/illegal and that the judges stay out it and focus on adjudicating alleged infractions.

Once we get enough experience with these systems to know what they really do (vs what the vendor claims they will do), then if data from the system shows up in a hearing, we can make reasonable judgements.