SRL Archive - Research & Development Lounge

I've been collecting data about the balloon random for some time now, but it doesn't come often so i'll need your help to get more results.

The principle behind is to calculate the average amount of pixel gathered from each balloon animal. From my past testings, it will be distinctively different for each animal so it can be used to identify the animal.

What you have to do is pretty simple, when you get into the random, simply pull the lever and run the following script:

program BalloonDebug;
{$DEFINE SMART}    //uncomment if not using SMART
{$I SRL/SRL.simba}

procedure DebugBalloon;
var
  i, TotalLength, AverageLength: Integer;
  TPA: TPointArray;
  TPALengths: array[1..64] of Integer;
begin
  ColorToleranceSpeed(2);
  SetColorSpeed2Modifiers(0.05,0.90);

  for i:=1 to 64 do
  begin
    FindColorsTolerance(TPA, 9902232, 119, 20, 387, 281, 22);
    TotalLength:= TotalLength + Length(TPA);
    TPALengths[i]:= Length(TPA);
    wait(618);
  end;

  AverageLength:= (TotalLength shr 6);
  writeln('Post the debug below AND what is the animal you are testing');
  writeln('Length array: ' + ToStr(TPALengths));
  writeln('Average length: ' + ToStr(AverageLength));
  OpenWebPage('http://villavu.com/forum/showthread.php?p=1113135');
end;

begin
  SetupSRL;
  DebugBalloon;
end.

When it's done, post the debug together with the animal being tested.
For the identity of animal, refer to
http://runescape.wikia.com/wiki/Balloon_animal
The cat and the dog may be slightly hard to distinguish, the main difference is their 'horns': the cat has the horns separated while the dog has them together.

Also note that the random will expire in 15min so once you get to it start testing asap. The testing should only take ~40sec for each animal so you have adequate time to complete it. For every random you can test up to 3 animals, but if possible, you can intentionally fail and hence test more.

Problem with this is that the amount of pixels per animal depend on the viewing plane of that animal..

Because of this, it will produce false positives and flag an animal as a different one.

I'm not sure this approach would work :S

From: Brandon
Problem with this is that the amount of pixels per animal depend on the viewing plane of that animal..

Because of this, it will produce false positives and flag an animal as a different one.

I'm not sure this approach would work :S

Yeah hence we will calculate the average length of pixels over 40s, which will even out the different views and the average will be distinct.
At the minimum it should achieve a high degree of accuracy, and since you have multiple trials it should be able to solve the random within 15min.

It's not worth trying, randoms will be most likely gone after EoC arrives.

From: litoris
It's not worth trying, randoms will be most likely gone after EoC arrives.

At the current state of botwatch (basically useless), i'm not sure if they are able to stop bots w/o randoms (not that randoms really stop bots), so they might change their mind. The biggest mistake you can make in RS is to believe what Jagex said.

From: riwu
At the current state of botwatch (basically useless), i'm not sure if they are able to stop bots w/o randoms (not that randoms really stop bots), so they might change their mind. The biggest mistake you can make in RS is to believe what Jagex said.

Your call anyway.

I've seen some sexi math from riwu, i believe in you!

Will do if I get the random!

If I happen to run across this random il do it :P. Ya should also look at the other accounts :P. Also how do you plan on detecting for the on screen balloons?

From: Ollybest
If I happen to run across this random il do it :P. Ya should also look at the other accounts :P. Also how do you plan on detecting for the on screen balloons?

Probably splitting the TPA of the animals by 1-2 spacing so that all of them can be isolated, i can double check the total length to ensure no overlapping. The size of box from TPA bounds may be used too. And if that's still not accurate enough, possibly checking for the concentration of points in a particular region within the TPA bound (but unless only checking the middle, the direction the animal is facing has to be considered). Could check TPA gaps as well?

You can stay in the jail as long as you want, but not pull the lever after 15min, so once the lever identification is done it's pretty easy to slowly test finding the animal.

Prison Pete loves me, i'll get you a account in the random soon.

Lol the only problem with this random, is that if you get it, and wait too long, you are released anyway, thus failing the random. Sucks, cause then leaving one over night wont do. Hey Riwu!!!

From: Rezozo
Lol the only problem with this random, is that if you get it, and wait too long, you are released anyway, thus failing the random. Sucks, cause then leaving one over night wont do. Hey Riwu!!!

Yeah that's one of the reasons why this random isn't solved yet. The only practical way is to get the users do the debug themselves when they get the randoms, as it's hard to contact random devs within 15min.

Lol, exactly. And I don't have that kind of skill…Or attention span at least :P