That polar was generated from GPS data collected in one afternoon of racing, but can't remember if it was one or two races. It took a while to get something that looked "clean"; tacks and gybes were removed, pre-start and post-finish was removed, messy bits were removed. Then I adjusted the wind angle to "balance" both sides of the plot.

I've thought about how to generate a cleaner polar plot from GPS data and came up with the following method.

- Pick a time close to high or low tide.
- Head out to a big open area with a fairly steady breeze and conditions.
- Go head to wind and measure the wind speed.
- Start recording with the GPS (or use this point as the beginning of you polar generation "run")
- Bear away and get up to speed on a close hauled pinching course.
- Bear away again, reasonably steadily, and work hard to obtain your maximum speed as you're bearing away.
- Keep bearing away until you gybe, then round back up again to close hauled pinching.
- Go head to wind and measure the wind speed again. (don't worry about direction too much, just get it from the plot when you balance it)
- This should take about 2-5 mins depending on conditions.
- Use only this loop to generate the polar.
- Take more time when coming up from a reach to close hauled because you will bring "latent" apparent wind speed with you.
- Setting the kite will change this procedure a bit.
- Do it again, maybe in the other direction so it is more of an experimental process and is repeatable.
- Change some settings and repeat, including measuring the wind speed before and after the loops.

When it comes time to process the data on the computer you will easily see the times you measured the wind speed, set the kite, etc.

That is probably a better methodology than just collecting heaps of GPS race data.