Oh if only GPS accuracy were as good as when you're driving in your truck, how that would make my life easy. Well, it's kind of a good thing that isn't the case...otherwise I might be out of a job.

Part of the problem is the best consumer GPS has a 10Hz update rate, and most are down in the 5Hz range. Turn that GPS mode to battery saver like I do and it's probably closer to 1Hz. That means at best you are probing the GPS receiver once every 10 seconds for data. Hey that's pretty good right? Nah, it's ****. At 25 kts you're covering 43 ft/s or about 4.3 ft every GPS parse. That's not terrible and within the realm of P1 GPS accuracy, but just about. What happens in the simple case of "my mast blocked the GPS signal", so you're loosing a packet of data ever other hit. That means now your sampling at 5 times per second, and traveling 9 ft between samples. That's right on the published 95% accuracy spec of 3 meters. And really you only get that accuracy with high PDOP values and when you're stationary. In short, if you're off by 9 feet at every reading for a 5Hz GPS, you could be off by as much as 18 feet between readings, 18 ft/0.4 seconds between readings = 45 ft/s or 26.6 kts!! That's the delta on your actual speed. Ouch right? Well, luckily the filtering within your receiver is usually much better, as is the accuracy (WAAS). But you can see how a small error between readings can result in a max speed burst several knots higher than your average. This is ignoring the effects of accelerating hands, masts, changing headings etc. that can all mess with the receiver itself (well, namely in degrading positional accuracy).

All of this can be eliminated by adding additional sensing equipment. For instance, 6-DOF IMU's do a really nice job at short term position prediction and hence velocity, can be sampled at high rates, well known filters etc. can be applied and you'll get good data. In conjunction with GPS information you are now looking at some really nice data. This is how UAV's manage speed so nicely (well, most also have an air data source, or speedo in our boat example). It's also how your phone knows it's orientation, the AC boats know their heel, and how those sick on screen VMG readings are produced with a high degree of confidence. The beauty is all this hardware is less than $300.00, if you just care about logging the data and getting within 1/10th of a knot.

Besides all this, average speeds matter much more. I bet you had much nicer VMG's on your I20 than the Solcat 18. To add a little more to that fray, most of the boats you owned aren't faster on a reach than your old solcat, more stable maybe so you can keep a higher average but faster? Probably not, the sail plans aren't optimized. I'll say with some confidence the new F18 hull shapes + rigs are faster on a reach than the I20 or pretty much any straight boarded 2-person production cat.