Page 1 of 1
Posted:
20 Feb 2003 20:25by cubenviper
may we ask to vsk designers to allow narrower crosses, which would be more spectacular and more efficient? i am not yet used to stay so far away(in dial ups for example)and get really disturbed by this. your opinion? :beer:
Posted:
21 Feb 2003 15:44by admiral 1
dial up
Port tack boat (P) and a Starboard tack boat (S) conferging on the same course coming abreast.
For all purposes only considering a situation with R10, R13 and R 16 are playing a role.
Based on R10 P has to stay clear, in such a way that S can change course without making immediate contact. The room S should have is then determined by the speed at wich S can change course and where it can go from there, within a timeframe that encompasses immediate.
Based on R 16 S can change course but must give P room to stay clear. P must be able to change course without making immediate contact.
When S changes course P gets a pen when immediate is over and a collison occurs, or for that matter can no longer be avoided.
stay clear
P may after immediate not be in an area where S can go within immediate. Even when in hindsight there would have been no contact ...
room to keep clear
S gets a Pen as soon as she enters the area P needed to change course to stay clear (at the time S started to change course) within immediate after S'es course change.
? define immediate: basically in real world measured by the commitee as they feel it might've been. Or a judge's feeling of what goes or might not go. Wind and water dependant. Mostly looked at only if someone protests. On the water without a commitee or a judge the more assertive skipper has most room, sometimes visibly comitting professional fouls... and then there's a basic 'understanding' of what goes and what doesn't ..... adopted by the sailors, the judges and or commitee. That can differ quite a lot from place to place, no matter how hard tried to define the rules explicitly.
? define course change: any eg. 0.5 degrees/ sec? don't even have to move the rudder to do that. 0.0000001/sec degrees would make holding course impossible. Course difference based on an average course over a certain time interval?
Any clue what the VSK judge is doing?
Objective: define parameters to control R10 and or R16 that will allow a dialup as you see it, and also a standard P vs S situation.
Make sure the algorithm (logic) is small enough to be computed every 50 ms? (framerate of 20/sec)
Anybody?
Posted:
25 Feb 2003 20:15by CAN Knot
may we ask to vsk designers to allow narrower crosses,
I'm beginning to think that it is not how close you are when you try to cross another boat, but how much you move your rudder to do it. You can dip a ROW boat by one length and still get a pen if you put your rudder hard over to dip. On the other hand, you can dip a ROW boat barely missing their stern and not get a penalty if you only make minor course corrections.
Any clue what the VSK judge is doing?
I think I'm starting to understand it.
Rules 10-13: Not much to get wrong here and VSK doesn't seem to have a problem with any of these rules.
Rule 14: No collision ever causes damage so no boat is ever given a penalty (as far as I can tell) for breaking rule 14.
Rule 15: Takes effect when a boat gains ROW when it is within two lengths of another boat, or when a ROW boat sails within two lengths of a keep clear boat (even though the ROW may have been established long before). I think this is the cause of most of the problems between port and starboard at the top mark. VSK will apply rule 15 to a starboard tack boat when it gets close to a port tack boat, regardless of when S really got ROW by tacking onto starboard. At times I've actually had R15 flashing red as I sailed all the way through the zone and dipped the port tack boat...
Rule 16: VSK seems to follow the ISAF definition of changing course: a boat changes course when it's compass heading changes. The trick is determining when a course change causse rule 16 to apply. And this seems to depend on three things:
1) Distance. The close the boats are the less of a change is required to invoke rule 16. VSK doesn't seem to pay attention to boats more than two boat lengths apart.
2) How fast the boats are getting closer together (i.e. boats on opposite tacks are getting closer much faster that boats on the same tack).
3) Total change in compass heading
The length of time that rule 16 applies seems to be a function of all three variables. The closer the ROW boat is to the give away boat when it changes course the longer VSK applies rule 16 to the ROW boat. The more the ROW boat changes course the longer rule 16 will be applied. If the boats are close and the ROW boat does a big head up or bear away R.16 might apply for several (crucial) seconds.
Rule 17: VSK seems to display this rule at the appropriate times, but I have no idea if penalites are applied appropriately.
Rule 18: is almost never applied. I've only seen R.18 in the ISAF window twice, and both times for just a split second. While we may give each other room it's likely worthwhile to keep in mind that VSK isn't paying any attention to R.18. A keep clear boat should keep well clear; I know I've been rear-ended while rounding the separation mark clear ahead and gotten a penalty for it... which really sucks. Similarlily, at the bottom mark VSK always gives ROW to the outside, leeward boat, rather than the inside boat. It almost seems that VSK has it backwards and gives precedence to the rules in parts A and B and only applies rules in C if they don't contradict.
Posted:
26 Feb 2003 11:55by admiral 1
It almost seems that VSK has it backwards and gives precedence to the rules in parts A and B and only applies rules in C if they don't contradict.
That just might be it.... then again I've seen rule 18 in the isaf window, even manage to get penalised on them.
sofar:
18.2 (a, b and d)
18.3 (a)
will have to get a second cd to test a little further as split screen doesn't hack it.
please keep in mind that as ruling evolves in split seconds around marks LAG may be a determining factor here.......
Posted:
26 Feb 2003 12:06by euphoria
will have to get a second cd to test a little further as split screen doesn't hack it.
With the NoCD crack from the Tools page, you don't need another CD. If you actually was thinking on a serial, well why not set up a LAN game as you can run multiple PC's on the same serial on those? I've done tests that way on 2 of my PC's.
Harald
Posted:
26 Feb 2003 12:21by CAN Knot
please keep in mind that as ruling evolves in split seconds around marks LAG may be a determining factor here.......
Most of my "testing" was done playing against the computer AI so lag wasn't an issue. (I figure the VSK judge will make the same calls regardless of who controls the boats). I've never been able to get R.18 to appear in the ISAF window under controlled circumstances; the two times I did briefly see it was during online racing. Maybe you need lots of lag before the judges will apply R.18???
Posted:
26 Feb 2003 16:51by admiral 1
With the NoCD crack from the Tools page, you don't need another CD.
need two key-codes to play online through the vsk servers (I hope at least so much was reckoned with)
PLaying against AI is a bit awkward as you can't control the AI boat..
Split screen is awkward controlling, and the display is somewhat distorted..
Posted:
27 Feb 2003 08:43by euphoria
Why not set up a LAN race?
Harald
Posted:
27 Feb 2003 16:15by admiral 1
Lan race is good to see if algorthms/parameters are okay.
I-Net is needed to see what LAG distorts.
I think it won't do much good if judge works correctly only in LAN or split-screen.
Posted:
28 Feb 2003 14:46by euphoria
I agree with the lag, but do you think you will have more lag in an internet race between two local computers than a lan race? Doesn't the data go directly from pc to pc, without passing through the duran server?
Anyway, there are many skippers in this pool that would be positive to join some sessions for the purpose of pen system testing. But information on how what the current algorithms are should really be received directly from Duran instead of us wasting our time.
Harald
Posted:
02 Mar 2003 16:53by admiral 1
Doesn't the data go directly from pc to pc, without passing through the duran server?
Thinking more like having a game between two pc sitting side by side, each with a dedicated internet connection, aswell as being able to play LAN. For the first a 2.nd key-code is needed to
get to the vsk server.
I also need it to have 2 pc with shared connection join a 8 boat fleetrace in the internet. Wich I hope will lead to the ultimate teamracing. Like the local sailing club throw a lan party racing against another sailing club. (with on both sides a case of beer in the middle ofcourse)
But information on how what the current algorithms are should really be received directly from Duran instead of us wasting our time.
That was what was concluded multiple times during beta testing. I'm sure Murray brought it up, i others brought it up and I know i did. Allas no matter NDA no word about governing algorithms. I can understand cause that will be half the net value in the game engine if not all what it's worth.
So no matter whether VSK lay open the algorithms, it must be possible to develop algortithms, based on parameter and object
description.
This way would also cause for more creativity and paralell thinking. It just needs enough critical brain mass. 
I'm sure vsk picked up ideas already from the various discussions. At least going throug the beta's and rv's (release candidate/version ) i conclude they are reading and listening allright.