Zezo forum:

Post by jfloviou | 2020-11-23 | 17:35:37

Hello,

Am I wrong or VR Winds have been wrong compared to Zezo for the past 1 hour ? I was thinking GFS download was not finished, but now it is, and still quite different.

23 Comments

Post by BooBill | 2020-11-23 | 17:42:33

Not sure where you are, jfloviou doesn't correspond to a boat name. However, if you are in the transition of the cold front where I am, just a few minutes boat movement can be the difference. The 2-3 minute delay with VR updating the position and then another minute or two to enter it in Zezo causes it to think you are in a slightly different position than where VR thinks you are. Also, rounding error in the interpolation of the GRIBs can be significant.

If that's the case, manual steering is going to be in order. Favour straight bearings over TWA. Try to make sure you are pointing in the right direction.

Post by NtsiabakaParadize | 2020-11-23 | 17:46:38

Same issue. The wind is not accurate. My boat name is ParadizeAgro

Post by NtsiabakaParadize | 2020-11-23 | 18:03:37

UPDATE : fixed

Post by zezo | 2020-11-23 | 19:14:17

It's expected in such conditions. See here:

http://zezo.org/forum.pl?tid=8194

Post by jfloviou | 2020-11-23 | 20:24:14

Thx Cvetan, but it is not exactly that if i'm not wrong: even VR Dashboard and speed&angles are wrong compared to Zezo: i guess the values in VR dashboard are calculated from server side values? I think Boobill might be right: interpolation and time delay, in a very light and shifty terrain ?

Post by zezo | 2020-11-23 | 20:31:59

There are three different sets of numbers, I don't say the router matches anything during that time. It could probably match the client, but that would be wrong because it still won't match the server.

Post by zezo | 2020-11-23 | 20:39:06

Here is how the difference between client and server looks like. Not during update, just around the front.

Boat is sailing fixed TWA. On refresh jumps in position occur because the client does not think it us where the server thinks it is.

One of the differences is the time interval. Server probably moves your boat 10 minutes at a time, using the same wind. Client moves it every minute, using wind interpolated down to the minute.

Post by BooBill | 2020-11-24 | 04:25:08

On the 00Z update the game definitely has a corrupt GRIB. Current matches Zezo and Windy. Then predicted path has a 60 degree lift in a few hours. Sailing a twa that should take me south, has predicted path swinging all the way east. It's like the front we just passed is jumping south and west. That's not meteorologically possible.

Post by karriv | 2020-11-24 | 05:42:57

Quite a difference server vs. client vs. zezo.

Dashboard reported interesting wind information, TWD varying every ten minutes between 23 and 36 since 04:00 UTC, e.g.

04:40 TWD 26.9
04:50 TWD 36.1
05:00 TWD 23.3

At the same time, zezo was showing 20-23 and the client was showing 36-38...

Makes life interesting. Delta between zezo and the game gives a flavor of real life, delta between the server and client? Perhaps broken instruments in shifty conditions?

Post by BooBill | 2020-11-24 | 05:45:29

Make sense of this. All these boats are within 2 miles of each other.

⎈ MONICK 24:30:00 EST 33195 19952.09 -8.9 80.6� LJ (Auto) racing 26�03'32.07"S 28�48'08.00"W 140.1 54.7 6.5 6.47 1.0030 0% Full Pack
⎈ wiizz 24:30:00 EST 34789 19954.98 -7.4 66.5� LJ (Auto) racing 26�02'01.13"S 28�50'18.68"W 140.7 55.3 6.4 6.36 1.0030 0% Full Pack
⎈ Big Bird @zezo.org 24:30:00 EST 34206 19953.96 -7.2 72.4� LJ (Auto) racing 26�02'48.40"S 28�50'14.42"W 178.2 50.2 7.5 7.32 1.0000 --- ---
⎈ machaon 24:30:00 EST - (4596.3) -5.9 92.8� C0 (Auto) racing 26�05'16.45"S 28�51'23.18"W 170.5 81.0 6.6 7.68 1.0030 ? ?

Cvetan, Wizz is pretty much on top of you and he has wind direction 43 degs less than you.

Post by zezo | 2020-11-24 | 05:54:45

That's probably the triple interpolation between 18z+6, 18z+12 and 0z+6 playing tricks on us. The front is in different (delayed) position in 0z relative to the 18z forecast and appears to be moving backwards in time.

Post by jfloviou | 2020-11-24 | 12:26:23

I do confirm issue still remain. Sailgrib with simple GFS file is ok, but Zezo is not. weird. Exemple:
position 26�54'39.99"S 28�42'32.00"W at 12h10UTC
Wind GFS (sailgrib): 126�, 11.2kts
Wind VR (as seen in VR dashboard: HDG 183.7, TWA 56.8 => TWD= 126.9 TWS=11.12
Wind Zezo: HDG 183 - TWA 60: TWD=123�. But displayed wind is right!
=> It seems zezo is not using proper wind for its current calculation, but displays the correct wind.

Post by zezo | 2020-11-24 | 12:54:32

If you see it in the display wind then the router uses it.

Thing is that the numbers (HDG, TWA, TWS) don't always add to 0 in the router.

The HDG is averaged a bit to avoid small fluctuations. Otherwise with the interpolated winds there would be a black dot every 10 minutes.

It's up to you to decide what to follow in any given moment - TWA or HDG. Here the polars come into play. Flat spots, sail changes etc.

Post by jfloviou | 2020-11-24 | 12:57:11

Ok Thanks. I had the impression this is a new behavior. I did not experiment such gaps before.

Post by zezo | 2020-11-24 | 13:10:45

And you were correct. I double-checked, and yes, the router has been using old data for a while. Fixed now.

Thanks for the observation. Would have taken me a bit more to notice. Do not hesitate to drop me a line if you notice something.

Post by PML | 2020-11-24 | 13:06:55

I concur with jfloviou: the red / green lines do not seem consistent with the routed path in terms of HDG / TWA currently.

Post by PML | 2020-11-24 | 13:09:04

Well HDG seems correct, but not TWA.

Post by YourMomSA | 2020-11-24 | 16:01:49

I'm curious about exactly what was wrong in Zezo... which wind timing were using what old data? And was there actually a problem in VR also, or not?

I ask simply because I went to bed about 10 hours ago, thinking I was in a great position to possibly be the leader at the Cape of Good Hope, and now I'm stuck in light winds and thinking much more pessimistically. It may just be that the forecast changed naturally out of my favor. That's fine. And if it was a mistake in Zezo, that's fine too. I'm not asking for the purpose of complaining... Just trying to understand what happened.

Post by zezo | 2020-11-24 | 16:11:52

About 6 hours ago I broke the update process a bit.

As result of it the 0z+12 was not replaced with 6z+6 at 10:30 UTC. Probably the forecast past +5 day was not updated either. I fixed it at 13:00 UTC

Right now the NOAA is a bit delayed.

I'm not aware of other problems.

If it was 10 hours ago then it's probably just the forecast.

Post by YourMomSA | 2020-11-24 | 16:42:45

Thanks for clarifying. Just plain bad luck, then. It could be worse. All plans in light and shifty winds are tenuously built on houses of cards.

Post by BooBill | 2020-11-24 | 18:13:13

My conclusion of what happened with the 00Z update (last night for us in NA) was that 00Z had the frontal line slow, so that it was behind where it was forecast to be in prior 18Z update.

For those of us that had just crossed the line, it appeared in the game to be a meteorological impossibility, a front backing up. Artifacts in the interpolation caused wind shifts of 45 degrees from one 10 minute refresh to the next. Zezo missed all this, since it used the old forecast for the first hour of the update and by the time it switched to the new, the issued had passed and everything synced up.

While a front backing up can't happen in real life, because we are sailing between forecasts of future wind, it can. If the progression of a front slows and you have reached the point where it was earlier predicted to be, it would appear to back up and you have to cross it again.

Not sure what your experience was over where you are YM, but for the tight pack I'm in it was a tonne of fun that kept me up till 2am. Some team mates who had gone to bed before the 00Z update started with TWA set, found themselves 100 miles off course this morning.

Post by YourMomSA | 2020-11-24 | 18:44:19

My experience was that the 4.2-4.8 kts that I should have been in for 3-4 hours lasted more like 8-10 hours. Which of course compounds itself because every mile made south would have been into more breeze. That's consistent with your theory of the front slowing down, I think. If you look at Leszek2011, I was pretty confident 14 hours ago that I'd be crossing him today. Looks unlikely now. Harder to tell how I'm doing against the giant pile of boats you're in. Probably all irrelevant in the long run... We'll probably all stack up somewhere a month from now.

Post by BooBill | 2020-11-24 | 19:17:26

Doesn't sound like there was anything you could have done about it. If the wind ahead of you evaporates on the next forecast, that's just tough luck. At least you weren't swung 90 degrees off course.