GPS Logging for and by Beginners
Well, the way I do it. Before I get into this I am a raw beginner at this. There are people who do this much better than me. Anyone from an AC or other professional team who reads this will doubtless do so with a great deal of amusement. Seek out other and better sources if you can, and take this with a considerable pinch of salt. Nevertheless I do find what I've learned from doing this helpful, and it requires no more than a smartphone and some free software.
The data and analysis we have here is from on a single day's racing in my International Canoe in light conditions on a moderate size reservoir inland. The boat is basically one of the old one design Nethercotts, but lightened and with some tweaks in the stern area. The exercise needs to be repeated time and again.
I have a Sony smartphone which is waterproof. I've installed a free app on it called GPSLogger for android, but there are lots of similar. I start that up before I go out, and shut it down when I finish sailing. The likes of a Velocitek speed puck will produce exactly the same sort of trace: the limits are I think mainly on the satellites, not the local electronics. Best results come on open water, worst results on narrow lakes surrounded by mountains.
OK, so what that gives is a list of points. It will also probably give you a maximum speed and a distance travelled, but these are of dubious reliability as I'll discuss.
It's probably possible to get an app that does all the analysis on a smart phone, even more likely on an Apple device. I haven't found an open source one I like. Those who are prepared to shell out will doubtless be able to get better. What I use is a program called GPSAR (GPS Action replay), which runs on Windows but is distinctly geeky to manage. That doesn't bother me, but I am sure there are more user friendly options, not least the same author's paid for option.
Anyway, load that into a display app and it will give you a trace, which might look something like this. It's probably possible to put a map and a satellite photo under it, but who needs it. Basically it's an impossible mess, but one can begin to see where the beat must have been, and also that it appears I didn't switch off the logger until I got to the Sailing club gate that day.
The first thing to rememember is the tracking has its limitations. Look at Fig 2.
As you can see from the scale during this trace the boat wandered about, very slowly, in an area of about 10metres square. So what? Well this trace is the time when the boat was tied up alongside the jetty (shroud tied firmly to the jetty, not moving about on the end of a long painter) while I had a cup of coffee so this this track indicates the limits of reliability. Occasionally you will get points that are much farther out. More sophisticated systems, which I imagine the the pro race tracking uses, can improve this, but this is what's available to the likes of you and me.
OK, What can we do next. Different software has different capabilities, but I'll run through some of the options. The first thing to do is to take out the extraneous parts of the trace.
Here I've taken out all the before and between race parts of the tracks, leaving just three races. They were sailed on similar but not identical courses: All quadrilaterals with the same start and windward mark, but different wing marks. The wind swung round between races 2 and 3, so what was a very biased beat became a straight one. This will mess up any analysis that includes wind direction, so it would be better to split them up and analyse separately. So lets look at just races 2 and 3.
This is clear enough that we can start looking at it like a sailor. The wind direction, by the way, is calculated by the software making some intelligent guesses from what look like tacks. You can see some dodgy tacks, you can see changes in heading caused by shifts, and you can see that the wind was still swinging in general trend. You can also see the speed of the boat by the track colours, and how much faster cross wind is than either reaching or running.
So let's get the software to do some analysis. The first thing everyone thinks of is speed. This software will calculate the best speeds over various distances, and we get a table like this.
And you go to the top of the table and see Best speed 14 knots, and say to yourself wow, and in only a Force 2. But it's nonsense. Look again, 7.3 metres in 1 second, but remember in fig 2 that adjacent plots can be some metres out, and this is just a measurement between two plots. Meaningless. Basically the longer the run the more accurate it will be. Of course there are other issues too. The longest leg on the race shown was about 300 metres, so the 500m runs aren't very meaningful either.
This section was built with GPSActionReplay. Visit www.gpsactionreplay.com
5 bests' average = 23.65km/h [12.77Knots]
Best Speed n°1 = 26.33km/h [14.21Knots] (7.3 meters in 1.0 sec.)
Best Speed n°2 = 24.33km/h [13.14Knots] (6.8 meters in 1.0 sec.)
Best Speed n°3 = 23.49km/h [12.69Knots] (6.5 meters in 1.0 sec.)
Best Speed n°4 = 22.21km/h [11.99Knots] (6.2 meters in 1.0 sec.)
Best Speed n°5 = 21.89km/h [11.82Knots] (6.1 meters in 1.0 sec.)
Best Speed n°6 = 21.86km/h [11.8Knots] (6.1 meters in 1.0 sec.)
Best Speed n°7 = 21.69km/h [11.71Knots] (6.0 meters in 1.0 sec.)
Best Speed n°8 = 21.47km/h [11.59Knots] (6.0 meters in 1.0 sec.)
Best Speed n°9 = 21.17km/h [11.43Knots] (5.9 meters in 1.0 sec.)
Best Speed n°10 = 21.08km/h [11.38Knots] (5.9 meters in 1.0 sec.)5 best 50 meter (at least) average = 14.28km/h [7.71Knots]
50 meter run n°1 = 16.15km/h [8.72Knots] (53.8 m. in 12.0 s.)
50 meter run n°2 = 14.2km/h [7.67Knots] (55.2 m. in 14.0 s.)
50 meter run n°3 = 14.09km/h [7.61Knots] (50.9 m. in 13.0 s.)
50 meter run n°4 = 13.72km/h [7.41Knots] (53.4 m. in 14.0 s.)
50 meter run n°5 = 13.24km/h [7.15Knots] (51.5 m. in 14.0 s.)5 best 100 meter (at least) average = 13.06km/h [7.05Knots]
100 meter run n°1 = 14.59km/h [7.88Knots] (105.4 m. in 26.0 s.)
100 meter run n°2 = 13.72km/h [7.41Knots] (106.7 m. in 28.0 s.)
100 meter run n°3 = 12.61km/h [6.81Knots] (105.1 m. in 30.0 s.)
100 meter run n°4 = 12.26km/h [6.62Knots] (105.6 m. in 31.0 s.)
100 meter run n°5 = 12.11km/h [6.54Knots] (104.3 m. in 31.0 s.)5 best 200 meter (at least) average = 11.95km/h [6.45Knots]
200 meter run n°1 = 13.47km/h [7.27Knots] (205.8 m. in 55.0 s.)
200 meter run n°2 = 12.42km/h [6.71Knots] (200.1 m. in 58.0 s.)
200 meter run n°3 = 11.77km/h [6.35Knots] (205.9 m. in 63.0 s.)
200 meter run n°4 = 11.14km/h [6.01Knots] (204.2 m. in 66.0 s.)
200 meter run n°5 = 10.97km/h [5.93Knots] (204.2 m. in 67.0 s.)5 best 500 meter (at least) average = 10.27km/h [5.55Knots]
500 meter run n°1 = 11.1km/h [5.99Knots] (502.6 m. in 163.0 s.)
500 meter run n°2 = 10.85km/h [5.86Knots] (503.5 m. in 167.0 s.)
500 meter run n°3 = 10.28km/h [5.55Knots] (502.4 m. in 176.0 s.)
500 meter run n°4 = 9.9km/h [5.34Knots] (500.3 m. in 182.0 s.)
500 meter run n°5 = 9.22km/h [4.98Knots] (502.1 m. in 196.0 s.)
What else. Well this software can generate polar diagrams.
These again can be misleading. An interesting thing about this one is that its so asymmetrical. A lot of this is an artefact of the course and the wind changes. The software makes an average guess of the wind direction and runs with that, so that's why there are nonsenses like the boat apparently sailing directly into the wind. That was just a big shift. In that race pretty much all of the runs were on one gybe, so there is little data for the other one. Similarly there were two reaches, one a bit shy of 90 degrees, and one a bit deeper. In this graphic the density of lines represents the number of points, and you can pretty much figure out the different legs! It shows that a race trace like this isn't that useful for analysing raw performance, because for so much of the circle there is very little data, and most likely what there is are points taken mid manouver where speed will naturally be low.
The dotted lines on this chart show vmg, and this is key when considering tuning details. You can see how low you can go before vmg drops off, and how rapidly it drops off as you sail hotter angles. I'm sure you can learn more about this with structured sailing. I did one race (not shown here) where I deliberately sailed what I thought were fast hot angles on one gybe, and deep but slow on the other. It turned out the deep but slow was getting me to the mark faster, which means that I'm now looking at optimising the boat setup to sail more efficiently at deep angles rather than hot angles.
So to try and get a bit more out of the polar I've done a bit of crude editing to make it symmetrical and drawn on where I reckon the actual polar might be, plus lines to show max vmg upwind and downwind. These are guesses of course. Educated guesses perhaps, but guesses. They also apply only for the light winds race this was taken from.
So what do we see. From the top it appears that it doesn't really pay to pinch the IC as hard as it can go, because the performance drops off a cliff, but it also goes off significantly if you go too low. That's not really a surprise, we mostly know that. The polar makes it look as if I should be tacking through 60 degrees. Don't dismiss this as proving the exercise is nonsense: go back to the traces and look at the wind changes. What this is telling us that in those conditions the combination of best angle *and* favourable windshifts meant that the best angle could be as high as 30 degrees to the average wind. There's also a bit of a suggestion that I was sailing on the right shift quite a lot, which is encouraging.
What else? Peak reaching speed is probably in the 110 degree region, but the trace is awfully sketchy around 90 degrees and from 120 to 150: it would be unwise to draw many conclusions.
The VMG (dotted line) really drops off a cliff both sides of the optimum running angle of about 160 degrees. I'm sure part of this is lack of data in the 140/150 degree region. However the sensational drop in performance as the boat goes towards a run is well worthy of study. From other traces I've learned that this peak vmg point, in these sorts of conditions, seems to be where the jib is flapping about and doing very little, but the main is still working properly. Once the main stalls out the boat halves in speed. So the optimum in these conditions seems to go as deep as you can without the main stalling, and to work on the jib setup so its still working that deep. Goosewinging with flow from leech to luff appears to offer promise. It also demonstrates that, much as I'd like to believe differently, you really do need that adjustable rigging so you can get the boom out further on the lee side. It really looks like a big gain.
Contact Me