Sisyphus: completed (maybe)

[Lotharsson]

Ergo all cars lose torque as they break down, just to varying values; the wheel number is fixed from the start though.

I thought I saw cars that did not lose any torque when pieces broke off. Will need to watch more carefully to verify or refute.

250.1 - PSO ran to 121 generations overnight. First posted code for a 250+!

250.3 - in sight of the end of the track!

eNqzf/fvcZ+ysbEDBwMYOIgWcjdGMF13YJ8d/57lzFn7Nxv9Vj9LS7P/uNTPe+fMWfZvry08FpaWav8p2FSuEsh/aHS0GER/+KGhnzlzloMQxBz7m3M8b282NrG/9+3mtvv//zswf/kdYQY058krn0tuaekObD3+Zd7GxkCljA5icXbpGkYJ9h/PKt1pAGtnd+C8kvQrrDDQ/qOxef2bM2eBYkz291sOffPYxG7/pVOOqS0tDSjG7MBZ+si6mVXY/nPAfcfLYPNY7W9Yeh+2k7a0f/nqgtdmsBiLg+it9i8Ji/7YP4l6X6gG1svowHNGSeLadjP7DzunvHwGFWPcLFxTH8pif6VqxSdjoN7D6zUZeO/tZziWvoFB3uYKA2tEL4N3oSPDikBGBo+94gx8LjsYnj6tZzg2bTZDIFMQw+vl2QwtR5IYVrrlMbRpyYHcDQAp4oiE

[Diseased]

Holy crap! I didn't even know that Sisyphus ever ended, I assumed the computer generated the curve ahead of you as long as you were still moving, or just stopped at some ludicrously high point. We need to breed a finisher.

As for the torque issue, my best guesses are: that the torque rounding system is 'optimistic' so that any decimal number, no matter how small, will round up to the next whole number (perhaps to prevent exceptionally small and light cars from becoming immobilized, although I've never seen torque 1 before and don't know whether it can be attained); that there is a mass density factor of the shapes that is has a wide range, so a reasonably large piece may not contribute to a significant drop in torque (assume standard number rounding); that because the mass of the shapes (assuming identical densities) would be proportional to their areas a small piece being removed may only make a negligible difference in torque due to exceptionally small mass; errors in your observations - or any combination of the those things. I have no idea if the mass density actually varies at all, I'm just trying to make common-sense conjectures.

It is also worth nothing that since the torque equation involves division by the number of wheels, the torque outputs have much smaller values for equally massive cars with more wheels and so equally massive pieces would be allowed to break off on many-wheel cars without the computer rounding down the torque number (e.g., if a four-wheel car has torque 127.96, displayed as 128, and loses a piece that drops the torque by 0.42 to 127.54, it would still display 128, but if a two wheel car with proportional torque 255.92, displayed 256, lost a piece of the same mass, it would drop proportionally by 0.84 and round down to display 255 assuming standard rounding).

[Faqir]

250.4

Code: Select all

eNqzv5W+1TFH/aADBwMY2K+cCQKz4Pxzguz2NnUzHdhDdByMXfbbf4CK74Sos98/2764smiTA1uz/mLNPB77YzvaT1kqbLW/dGb1ybuxsjDz7HduCBY7w1hk/zbr02qFV/fsDz+9pyfOMhloFKODIKOw3Lk4Bvsnjdda3d5mAMVY7G822Wfnny+xf55jGqBdyvgfCBw4bwk6vmK1tf9gw8S5wf4dUB2TA/+6vKSEdVvsjzQLed1h7AaKMdt/WzH9nuuSP/ZfbGLCei0vgPTaf+XrEj61Ldn+paOT6j7/cqA6dgehzS+cF99mtP/0KNguVqwWLCaqxXb9t62B/Z1mGx0hxQ4GfT4BhqLL6xmenHvOsOv8f4YwnnAGj/8ODAfUbzP8+1fB8OLIe4Z//xcyyE/5ydDwbCXD0eMtDB//3WDgyeNhOPlfHeRGALAgj7s=

[aggrazel]

Wow, nice job Faqir ... nice to see the track is technically winnable.

[Lotharsson]

250.4

Excellent That's interesting because I've also been trying a smaller-wheeled car for Sisyphus to see how far it can get. You can make it go quite high

[TechSY730]

250.4

Code: Select all

eNqzv5W+1TFH/aADBwMY2K+cCQKz4Pxzguz2NnUzHdhDdByMXfbbf4CK74Sos98/2764smiTA1uz/mLNPB77YzvaT1kqbLW/dGb1ybuxsjDz7HduCBY7w1hk/zbr02qFV/fsDz+9pyfOMhloFKODIKOw3Lk4Bvsnjdda3d5mAMVY7G822Wfnny+xf55jGqBdyvgfCBw4bwk6vmK1tf9gw8S5wf4dUB2TA/+6vKSEdVvsjzQLed1h7AaKMdt/WzH9nuuSP/ZfbGLCei0vgPTaf+XrEj61Ldn+paOT6j7/cqA6dgehzS+cF99mtP/0KNguVqwWLCaqxXb9t62B/Z1mGx0hxQ4GfT4BhqLL6xmenHvOsOv8f4YwnnAGj/8ODAfUbzP8+1fB8OLIe4Z//xcyyE/5ydDwbCXD0eMtDB//3WDgyeNhOPlfHeRGALAgj7s=

Mind if I "borrow" this and seed it into my current population?

[Diseased]

I've taken Fakir's and mine and seeded them into a 2-car generation PSO run. Predictably there was chaos at first, but PSO has the mechanics down and improvement (on the first successful run, not 250.4 of course) is happening fairly steadily, mainly through tiny permutations of the Fakir model which is what I was doing in designer for a while anyway. It occasionally makes hybrids, which are stable and reasonably good but far too high off the floor at the moment.

This is so exciting! I think that finishing Sisyphus will be this game's real Apollo 11 moment, get cracking everyone.

[Lotharsson]

Predictably there was chaos at first, but PSO has the mechanics down and improvement ... is happening fairly steadily, mainly through tiny permutations of the Fakir model which is what I was doing in designer for a while anyway.

I've been running some PSO on my 250.3, and after a while almost all of the cars score < 1. Not sure if this is how it's meant to be...

[Diseased]

I've been running some PSO on my 250.3, and after a while almost all of the cars score < 1. Not sure if this is how it's meant to be...

Attached: my 2-car generation graph. It's easy to judge very accurately what's going on with the vehicles with it using simple deduction, e.g.:

black line 1/2 way up = 1 good car, one failed car
black line rock bottom = 2 failed cars
black line top = two good cars

so I'd say that I'm easily getting over 50% worthwhile vehicles if I discount the starting period.

[Faqir]

Mind if I "borrow" this and seed it into my current population?

You can use it if u want to.