Difference in Routing Prediction?
Post by iconPost by magnusstrom | 2011-12-14 | 19:47:31

I pretty much missed the start of this race, and am down the pan, but managed a respectable 119 without any routing on leg 1.

I just started to use this routing, but there is one thing i struggle to understand about it is this:

Say I do a routing at 3pm, and it pretty much tells me to bang east at fast CO speeds. A few hours later, without a wind update having happened, you do another routing and it does a pretty fundamental shift in tactics with a much higher course on a jib. This has happened a couple of times. Surely the routing shouldn't change (unless there's been a wind update) as it should have anticipated where I should be those few hours later?

Is it stupid to do a routing update too often or how does this work? Not quite with it!

M

commenticon 9 Comments
Post by iconPost by Wildheart | 2011-12-14 | 20:12:07
ahhhh 6 hrs. Ty Zezo!
Post by iconPost by zezo | 2011-12-14 | 20:18:42
Sometimes there are two or more very close routes and small noise in the system can make the optimizer prefer one of them. One source of noise are the non-interpolated over time winds - real forecast would be smoother.

Another one is the bigger time step used for the longer forecasts. It generates some error, which is probably less than the weather uncertainty, but makes small changes in the star position generate huge differences in the routing.

It also depends on the weather systems in action. Wit a single stable low pressure system dominating the region it won't change its mind that often.

There are two ways to deal with the issue - 1) Check the forecast every 6 hours when new data is available, and then just execute the suggested track. 2) set custom destination at some point that prefers one of the tracks, according to your own strategy.
Post by iconPost by magnusstrom | 2011-12-14 | 20:22:09
Thanks. Need to play around with it more. VR tool still only works on PCs right.....?

StromArch
Post by iconPost by magnusstrom | 2011-12-14 | 20:25:28
@zezo i tried to figure out your route as a test, but it made no sense to me what you were doing on the 'water' with what regards to what SS was indicating... but it all makes a bit ore sense now! Ta.
Post by iconPost by zezo | 2011-12-14 | 21:08:57
I'm aiming around 10S 70E. like it better than closer to Madagascar. Does not look so good now, but it's too late to change my mind and will stick with that decision for a while more.
Post by iconPost by pgfiore | 2011-12-14 | 21:29:15
"1) Check the forecast every 6 hours when new data is available"

Doen't the updates happen every 12 hours? mornign and evening? To match with the VORG?
Post by iconPost by zezo | 2011-12-14 | 21:34:33
That's only the weather valid at the moment. The forecast you get at noon for the evening change is probably more accurate than the one you got in the morning.
Post by iconPost by pgfiore | 2011-12-14 | 22:43:28
oh I got it. Is the vorg the same?
Post by iconPost by zezo | 2011-12-14 | 23:02:00
No. It would be too hard to explain to users ;) So they don't update at noon and midnight. I use all possible data without disturbing what is used at the present moment.
border
Topics list
Posts
border
7
border
border
26
border
border
53
border
border
42
border
border
12
border
border
5
border
border
17
border
border
2
border
border
4
border
border
3
border
border
1
border
border
35
border
border
3
border
border
2
border
border
11
border
border
4
border
border
15
border
border
8
border
border
5
border
border
11
border
border
5
border
border
Copyright 2009 by ZEZO.ORG. All Rights Reserved.