How accurate is the Sailonline simulator?

Just how accurate is the ocean racing simulator run by Sailonline.org?

And just how well sailed are the courses?

Recently SOL finished a facsimile of Leg 1 in the 2011-12 Volvo Ocean Race . A few minor differences in the polar were noted, by the time SOL reached the Noronha mark – namely, sailonline were little faster in the light airs, while the real boats were a little faster in the heavy breezes – but this is somewhat expected as the SOL polar was modelled after the 2010 designed yachts. SOL uses the NOAA GFS gribs for it’s weather data during the race, updated according to NOAA’s schedule – four times daily. So, provided with accurate boat performance and weather data, how does SOL perform?

Answering this question normally, only a technical mathematical answer could really be given. But the recent running of the Volvo Ocean Race (VOR), parallel to a facsimile of the course and boats on SOL, the ’Sailonline World Race’ (SWR), offers a unique opportunity to get a ‘real world benchmark’ of just how accurate SOL is.

And just how good are those who compete there….?

Numbers still provide a good comparison though – and Elapsed Time is the obvious metric to compare. Team Telefónica, a Spanish team, won Leg 1 of the VOR in 21 days 5 hours and 14 minutes. Leg 1 of the SWR was won by the boat ’76Trombones’, sailing out of Canada, in 21 days 10 hours and 56 minutes. The time posted by 76Trombones, would have resulted in a 2nd place had a boat in the VOR set it.

Now, just reading those times, they are obviously close, but there is, in fact, only a shade over 1% between them. 1.1% more precisely. That is simply amazing, over 3+ weeks or more than 500 hours, of ocean racing.

Now also consider budget and effort. Technology etc. While it is clear that a lot of time and dedication went into creating the SOL simulator (and I refer almost exclusively to the server-side of things here), it rather pales when compared to multi-million dollar budgets and building on years of manufacture & design experience and development. Likewise the number of man-hours Telefónica must have dedicated to this race – even divided by 9 to compute a ‘per race’ figure – probably equals the entire life-time of the lone skipper/navigator of 76Trombones (yes I pulled that ‘lifetime’ figure out of nowhere, but it seems realistic to me).  I can hardly imagine more diametric opposites in both cases…

During the race the real boats had 11 crew – and, sure – a lot more to do than the virtual, ‘armchair’ equivalent, but even 1 of them dedicated to navigating and routing, still had 8 or more hours a day to dedicate to the matter, in between other duties on-board  - OH, as well as being a world class professional and highly paid for his/her services. This roughly equates to the navigator on 76Trombones dedicating 1/2 his days, 7 days a week, for over 3 weeks, to the race – as well as lead his, normal, non-SOL life (in which I have no doubt he is likewise of world class in chosen field)…. While not quite as opposite as multi-million dollar teams vs an individual, there is still an enormous gap.

Yet despite this, there was only a little over 1% between the times recorded by the 2 boats, one real, one virtual. I just have to say it again, that is awesome! Clearly, when the yacht performance and meteorological data is accurate, SOL provides a very accurate simulation of the real environment. And clearly, when competing there, not only are you acting within such a realistic environment, but you need to post a ‘world class’ performance to finish well…

One final note: There is an officially supported virtual simulation of VOR, the winning time on which was over 17 hours behind Telefónica (12+ behind 76Trombones). That’s close to, but SOL is 3 times closer! The VORG winning time was more than 3% outside the actual winning time. Indeed, over 80 boats on SOL beat the winning time on VORG, more than 6% of the SOL fleet! (VORG measures time quantized to 10 minutes)

It should also be mentioned that differences between the two simulations meant that for most of the race, the conditions were such that SOL should compute a slower average boatspeed – namely sailing into diminishing winds is faster, the longer the simulator assumes ‘constant weather’.

So: a serious ‘Well Done’ to Sailonline.org, 76Trombones …. and the top 80 finishers

Truly a world-class effort and result, achieved on a world-class ocean racing platform.