PDA

View Full Version : Is the size of a worm a random one?


evilworm2
17 Jun 2006, 22:14
Omg. Watch this.

It all happens in the very first move.
It seems that the size of a worm is beeing altered randomly.
I didn`t expand or reduce the rope. (Neither arrow up nor arrow down.)

Here´s the replay:
http://ez-files.net/download.php?file=f4f5fc73

bonz
17 Jun 2006, 22:52
Huh?
I didn't see anything.

If you mean that hyperspeed bumping between the 2 walls and the worm, you can clearly see and hear that the distances between the other worm and both walls are different, resulting in an optiacal illusion.

evilworm2
17 Jun 2006, 22:57
If you mean that hyperspeed bumping between the 2 walls and the worm, you can clearly see and hear that the distances between the other worm and both walls are different, resulting in an optiacal illusion.

I answer in german because i don´t know the english words and bonz is speaking german too.

Genau das meine ich. Aber warum bleibt der Wurm nicht auf z.B. der rechten Seite und fliegt durch den Wurm durch? Der sollte sich doch für eine Seite entscheiden? Für mich sieht das so aus, als ob der Gegner-Wurm mal grösser und mal kleiner wird.
Schau noch mal genau hin. Eigentlich sollte der Wurm am Seil die ganze Zeit an eine Seite des Gegner-Wurms bouncen. Nicht an beide.

CyberShadow
18 Jun 2006, 12:39
As you might have known, there's usually a speed limit on all object movements (16 pixels per frame for worms on ropes). If it were to be removed, bouncing between walls would accelerate you endlessly.

Now, in that particular situation, the worm was moving on the trajectory of a circle arc - the radius (the rope length) was constant and only the angle varied. If you were to check the points of the worm's coordinates each turn, you'd notice that they are indeed placed on a circle arc.

However, the movement BETWEEN frames is always linear. W:A traces a line between source (X1,Y1) to destination (X2,Y2), and checks collisions for every point on the line. Therefore, if the points on either side of the worm are far enough from the point where the arc passes through the worm's collision mask, the line between them might be far enough to go higher than the collision mask.

Bouncing between the wall and the worm varied these "jump" points - the coordinates of the worm at each frame. Therefore, you can expect this to happen with a predictable pattern - as long as the speed doesn't vary :)

Of course, finding this combination of angle/rope length/other object's position is another problem. The difference must have been less than a 100-th of a pixel - so, quite a find :)

[UFP]Ghost
18 Jun 2006, 14:10
As you might have known, there's usually a speed limit on all object movements (16 pixels per frame for worms on ropes). If it were to be removed, bouncing between walls would accelerate you endlessly.

Now, in that particular situation, the worm was moving on the trajectory of a circle arc - the radius (the rope length) was constant and only the angle varied. If you were to check the points of the worm's coordinates each turn, you'd notice that they are indeed placed on a circle arc.

However, the movement BETWEEN frames is always linear. W:A traces a line between source (X1,Y1) to destination (X2,Y2), and checks collisions for every point on the line. Therefore, if the points on either side of the worm are far enough from the point where the arc passes through the worm's collision mask, the line between them might be far enough to go higher than the collision mask.

Bouncing between the wall and the worm varied these "jump" points - the coordinates of the worm at each frame. Therefore, you can expect this to happen with a predictable pattern - as long as the speed doesn't vary :)

Of course, finding this combination of angle/rope length/other object's position is another problem. The difference must have been less than a 100-th of a pixel - so, quite a find :)

2 things:

1. wow, i'm ognna reas this again as i skimmed it and it looks smart,
2. Is it possible to
" (16 pixels per frame for worms on ropes). If it were to be removed, bouncing between walls would accelerate you endlessly."
remove that? cause i would like to play a game like that as that would be sooo cool. deadcode should make that an option, uber-fast roping.

CyberShadow
18 Jun 2006, 14:25
Ghost']2. Is it possible to
" (16 pixels per frame for worms on ropes). If it were to be removed, bouncing between walls would accelerate you endlessly."
remove that? cause i would like to play a game like that as that would be sooo cool. deadcode should make that an option, uber-fast roping.
This is already implemented in the test versions of the game, and will become a scheme option eventually.

[UFP]Ghost
18 Jun 2006, 15:11
goodie :D :D

[UFP]Ghost
18 Jun 2006, 15:12
can u re upload the replay?

evilworm2
18 Jun 2006, 15:13
As you might have known, there's usually a speed limit on all object movements (16 pixels per frame for worms on ropes).
<snip> A lot of words <snap>The difference must have been less than a 100-th of a pixel - so, quite a find :)

Thank you for your explanation. :)

evilworm2
18 Jun 2006, 15:16
Ghost']can u re upload the replay?

Why? The download link is still working.

[UFP]Ghost
18 Jun 2006, 19:17
not for me?

nvrm now it is lol, not before :p