Post by iconPost by ctr | 2014-04-07 | 09:15:50

Hello Cvetan
I found a gap of 10 minutes when the wind changes ...
I think on zezo, "bascules" take place at 8:05 am instead of 7:55

commenticon 6 Comments
Post by iconPost by zezo | 2014-04-07 | 09:49:05
This is a good question.

Yo can say tat there is a 10-minute difference and it affects not only the wind changes, but also programmed turns and all turns/tacks in general.

The problem is that the optimizer assumes that the game engine moves the boat at the 0 minute if the cycle, and we know that it does not.

Well, it tries to guess if your boat has moved in this cycle or not, so asking for a route at :01 and :09 (before/after the game cycle) should give the same result, but the logic is not perfect.

So what actually happens is this: let's say you have to make a turn after one 10-minute step. It's 15:01 now. You ask for a route and get the answer is "your heading at 15:10, which in my view is in 10 minutes should be xxx degrees". If you ask at 15:09 the answer is "At 15:10, which in my view is in 10 minutes ...." and that's where things get complicated.

It's hard to explain, but the general idea is that if the optimizer says "your boat should already be heading 112 at 15:10" that means that it should be heading that way at the previous game step, so if you program a turn you would set it at 15:06, or whatever that particular game offset is.

The moment of the wind change is especially complicated because the game will move your boat at 19:06 for 10 minutes at the old speed, and then you know what happens if you have auto TWA and programmer on at the same time - the game itself falls into similar problem. It would calculate your TWA on old direction and new wind and in general send your boat in a very wrong way.
Post by iconPost by Sir_Petrus | 2014-04-27 | 00:31:43
Dear all
I found this post very intersting, because I had some realy nasty problens with programing turns, according the advice of zezo.
I read very carefully, the explanation above but I was confused.
1st - the 10 min cycle gap. it is a back offset or a forword offset?
2nd - the exemple is not very clear for me. When refering wind chage, does it refer changing wind cells or, time change (8:00 and 20:00 CET)?

I had lots of problems in the frozen (3 min offset).
Problems like turning 1 cycle after, the correct time, and droping in the wrong cell.
Problens in changing cell, with TWA on, and programing a turn. Find the boat going back and loosing 4 cicles etc.
So I tried to figure out what was going wrong, and I start to program the turns, instead of the minute 10 as per zezo, in the minute 10-7, or lets say previous cicle, and I found thats the way it should be.
Based on that, I think the game program, only looks at our boat when it reaches the cycle time. doesn't matter if we make a program to turn 3 or 4 minutes before.
Also I belive that the moves goes with this order of actions when it reaches the time: 1st- move 2nd- TWA 3rd-your program.
Now I stil din't saw what happens in time wind chage or what happens in the middle of a cell.

Best Regards
Post by iconPost by zezo | 2014-04-27 | 11:24:45
By "wind change" above I mean the forecast update at 8AM/PM

And yes, it's very confusing, even for me, when I look at the source code.

I try to distinguish the case when you check the route before/after the game cycle, but that is not perfect because as we know the boats are not moved instantly, it takes up to 2 minutes, depending on ho many boats are in a race, so there is a period of time where the data is uncertain.

The way to look at it is to interpret "you have turn coming in 10 minutes" as "you have turn coming in 1 game cycle"

So if at :02 it says "you have turn in 10 minutes" you can actually execute the turn in 2 minutes, or somewhere in the next 10 minutes, but not later. I use the rule "10-7" or "previous game cycle" too and it seems to work.
Post by iconPost by muscadet77 | 2014-05-13 | 19:06:43
Dear Cvetan,

I have another question.
Until recently, the winds forecasts were updated on Zezo about one to two hours before the updating on VR. For instance at 4:30 pm or 5 pm instead of 6 pm.
Now it is at the same time as on VR, in this example at 6 pm.
Did it really change or am I mistaken ?
Best regards
Post by iconPost by zezo | 2014-05-13 | 22:19:22
The update times move one hour later with the change to daylight savings time (March 30th this year) because NOAA works on UTC.
Post by iconPost by muscadet77 | 2014-05-14 | 00:55:13
OK, thanks for the answer, I didn't know that :-)
Topics list
Copyright 2009 by ZEZO.ORG. All Rights Reserved.