Obsessive Compulse Tournament - Update 1
In case you haven’t heard of our AWESOME contest, link here. There’s still a ton of time to enter!
Wow, so we’ve been getting a LOT of entries… nearly 120 unique people have entered since last week! It’s been awesome to see everyone submitting the most elaborate and exciting level designs we have ever seen in Compulse, and you guys certainly seem to be getting better… some scores have improved by nearly 10-20 points in a few days. We haven’t OFFICIALLY checked any layouts, but I have to say that these scores I am getting are mind boggling.
We have several updates for you!
- Many of you have caught a minor glitch in the teleports. This is caused by the little circle going into the center of the teleports changing the teleport size, meaning the ball may or may not hit it as it pings. This glitch is most apparent in level 15 and 23, and may cause different patterns. The chance of these patterns differentiating is about 1 in 10, but it is still significant enough to warrant a warning. We acknowledge the problem and if your level works part of the time we will know and count it as okay… but make sure to drop us a note if you notice that one of your levels works part of the time (many of you are doing this, thanks). This problem is only on teleport levels.
- Yes, we know you can hack/slash/edit/modify/glitch the file. But it’s not going to make a difference since you are submitting your level designs to us. Please do not send us bogus scores from your awesome cheat engines.
- Some people are emailing us 10 times a day. Please send it once at the end of the day :(. Once or twice a day is fine but don’t spam my inbox!
- We cannot reveal the top scores to you. You have to use your better judgment (and listen to your rambling co-players in the main ArmorBlog thread) about what scores are floating out there.
So thanks for playing thus far! I’ll give you a hint as to where you may or may not be at the moment: there are 40 entries that are at or below the score of 80! A bit vague for you in the lead, but now you know at least where you are…. a tiny bit.
Remember you got until May 7th to get your best score! The original contest page can be found here.
[...] April 28 - Update #1 [...]
Yay one of the fist comments and one of the top 40!
Hey guys. Just to let you know that I have just sent a 62 score report. I think I’m pretty close to the “impossible to go lower” point.
I’ll keep trying.
Sog.
“This glitch is most apparent in level 15 and 23″
I noticed the teleporter glitch in lev23. In lev15,.. not so much. Perhaps if there was a teleporter in level 15 it would be more obvious.
I’ll put this one out in the open -> Would it be legal to ‘brute force’ solutions? (no, I haven’t been doing it.) It’s probably doable with some effort -> and you could probably speed it up if the game was decompiled to obtain the exact trajectory and hit-testing procedures and the code rewritten in something like C (inline brute-forcer, not a bot that drags the tiles!)……
That would be a fun script to write
“Some people are emailing us 10 times a day. Please send it once at the end of the day :(”
Aw. Sorry
It’s just that I think a few hours might make a difference, you know.
I had thought of brute forcing it as well. I tried it a few days ago looking for 1 tile soln to lev26. Decompiled, inserted automation code, turned framerate to max, recompiled and left it running. It took 3hr to do complete search. At this rate looking for 2 tile soln would take approximately 72yrs. Much faster in lower level language obviously, but porting to C, such that it is pixel perfect identical program would be nightmare.
Oh and on lev26 after searching all of the thirty thousand or so possible locations with a left/right tile, no solutions were found.
Wow. You sure are desperate XD
Thanks for letting us know
Na, it only took about 15min to write the 10lines or so of code that it took to automate, noticably less than the time I spent playing with it by hand. Pity I didn’t hit on the idea earlier.
Anyways, who can beat my score on ‘Sleepless Knight - assassin’. Best time is 192s.
What! You decompiled to bruteforce it? That’s just lame
I can’t believe you! You don’t even know if it’s legal!
Anyways, smart thing to do would be to give some constraints. It wouldn’t be hard to put the l/r tile in level 26 an upper bound as above to starting point, left and right bounds as so they aren’t in line with the ball… Woulda saved some time
I’ll bruteforce once I get the green light =P Which I hope to get.
Hmm, the only rule seems to be that solutions are valid, how you get those solutions is up to you:). Besides I shared the only result that I got.
And yeah for lv26 I bounded the x coordinate between 266-339. Interestingly if you are brute forcing there is another constraint to consider and thats how long you let it run before you give up. Its not uncommon in compulse to have it go forever without hitting a mine or reaching goal. For lv26 it was obvious when searching for 1tile soln, as after the 4-5th bounce there is no chance of ever reaching the goal. To model this my automation code aborted attempts once they went past 400 frames. For other levels this cutoff would be difficult to pinpoint.
Oh, and I don’t guarantee lv26 is unsolvable in 1tile. Sometimes decompile/recompile damages things. Also I only searched with integer values. As someone pointed out near the start of the comp you can actually use fractional pixel values in the ‘load’ window. Even after reading the code it is unclear to me whether fractional pixel values achieve anything, Flash 8 has such nonexistent type checking and i’m not intimately familiar with how the library functions behave with regards to any internal type casting… The exact internal behaviour of the library functions is another thing you might like to give serious consideration to as you are porting to C:).
Although I can be reasonably certain I am not in the running for any honours, and either way I played for the fun of it, if anyone who decompiled and brute force searched won, I’d be very disappointed, as it’s rather against the spirit of making efforts to discover solutions in your head and do the hard work of painstaking pixellated placements. The solutions you find, working, is half the reward
I did it all in my head. Hours of work. I’m proud of my score.
Yeah, I did all mine by hand too. I probably didn’t explain myself very clearly in the previous posts. My point was that I doubt it is possible to find anything useful by brute force, there are just too many combinations, even if you port to a more efficient language. The only exception is those levels with potential one tile solutions and if you are good you would have noticed relatively quickly that the only one of those that was of any interest was lv26.
Да уж, судя по комментариям тут на блоге, никому не повредит умение ярче высказвать свои мысли, вести интересную беседу, быть остроумным в конце концов…