Big Air

[Lotharsson]

I've also made it up to 421.1. It doesn't show any signs of making the last jump, though. Any thoughts on how to tune or breed for that?

I haven't spent much time looking at it, but the last jump just looks impossible at the moment. I have 2800+ generations that haven't improved on 421.1, so I'm not sure if evolution - at least the way I'm currently going about it - is going to find an answer.

Here's my car in case it gives you any bright ideas:

Code: Select all

eNqzfz7tk2Wdk4EDM7eP2brbefavlS48afmy2P7tQ5cVqyKP2X9gAAMH5qTbyaGvLsL5bEKcPXkC6fYrL62WfNSlYf9r+ftp1inp9m9+rjy7x+65/Ytu55nL1VTtHwqdcK3gL7I/eGz/7gnF9fYvdFT3H57Tav9DMXRK9YJpQKPYHUT/Mum/aVxl/4MBBpjtP4junxwwpQRJjMH+eFXUx5mTQu2veq/quBzAAlb39FSwcuOk//bfTbKe/bY8CxJzYPzg8t1mtbn9twWBgmrhsUAxRgfWhVd32zovtL8zodUpzmUuUIzFQciewfTIER37e/z85Xt8TgLFWB1YklYqneIsst/1cc3uNYVaDH5zvzGEm0syPF8+n6Gz5iZDeMw9hsTd7gw/nnQwnOiawZAfl8uw+E45g/d6B4ZdZf0MArdPMKw082OQjVjCcHKjJNBMJgAV/ooc

[ltsply2]

FWIW, Big Air is a very good example of the "local optima" problem. In order to make the last jump (which I believe is impossible) your starting pool needs to be everyone that successfully makes the penultimate jump. From there, the cars that can stay in the air the longest for the last jump are your best seeds. But that isn't the criteria the algorithm uses. If there is any car that beats 420.8 they automatically become the seed even if they are very bad in the real criteria necessary for making the last jump and taking that step forward...

https://secure.wikimedia.org/wikipedia/ ... _landscape

My suggestion for anyone trying to beat big air is to use the least stringent optimization criteria (genetic evolution, high mutation, no elite selection, roulette wheel) with a starting seed of a bunch of 420.8 cars that are as different as possible. Oh and this is a case where more generations is not likely better. I'd reset every couple hundred to your original starting seeds.

[envertigo]

I have a car that completes it. If I knew it was a big deal I would have made sure to save it. There is a chance I have it in a garage on another computer (aka notepad.exe)

I do have an idea of how I got there... start with a few hundred generations on speedway (or was it Sisyphus?) then let them shatter on big air for a while to toughen them up.


It's also possible I did some breeding... where I start a car on any track, next-next-next until I see a variation closer to what I want then start over with that as a seed.

(for some reason I am loathe to hand-tweak and haven't done so yet )

[Lotharsson]

I have a car that completes it. If I knew it was a big deal I would have made sure to save it.

Yes, it's a big deal, at least to folks here See if you can find it!

[envertigo]

Yes, it's a big deal, at least to folks here See if you can find it!

I think it's gone. But I pasted in the set that I started with. I believe I put this set on speedway for a while (~100 gens) then put the result back on big air. I am attempting to recreate all that here on the fast computer. I will post as soon as I get it.

These cars are somewhat immature but they have the right base.

Code: Select all

eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfAQTA+tly+X+WcCw0v5LFJfk7t27wWKi8/fMfX7lhf3HsC0mPzo6wGISq6sag/k07J/kf/zQ6ScJErN/7fn3Sd3fGPuTu1dtAroJrE5w84+Vr98K2L+5qenbc+YMQ4+KC8PTF88YDJavZki8UMEwJa6YwVXpDEPcVlGGgLZfDM657gzL5s5h0EptYtDjE2QojtJjeFvozjB3qy6Dm7MsyN0ANJuh0Q==
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTs9wUl3Xet+P0fCByYRPS91Hdp2H9dlXv61ykzoPGsDFBg/wPGYmB2EL4X6TVB8BFMDKyXLZf7ZwHDSvsvUVySu3fvBouJzt8z9/mVF/Yfw7aY/OjoAItJrK5qDObTsH+S//FDp58kSMz+teffJ3V/Y+xP7l61CegmsDrBzT9Wvn4rYP/mpqZvz5kzDEeObGB4+uIZg8Hy1QyJFyoYpsQVM7gqnWGI2yrKwHmamcE5152hd64ug0GwA4MenyBDcZQew9tCd4a5W3UZ3JxlQe4GANytncw=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0ODCl/zWt0Pps/4EBDOzfbok+Pr28wv6J3/qZOWfOOLDv7q9YeDXH/vLLrXeVjY0cWG65C38z0LN/yOXjUXLmjP0nLldVho4O+3uel/IXKSnZP39sbuP37t1/IHBgEtH3Ut+lYf91Ve7pX6fMgMaz2n+Y90hfcIql/Q8GGGB2EL4X6TVB8BFMDKyXLZf7ZwHDSvsPp39mbhQUAIk7iM7fM/f5lRf2H8O2mPzo6ACrk1hd1RjMp2H/JP/jh04/SZCY/WvPv0/q/sbYHwwJmfT40F+wOrH6Ob73o9Tt39zU9O05c4ahR8WF4emLZwz6EjwMiRcqGKbEFTO4Kp1hiNsqyhDQ9ovBOdedYdncOQxaqU0ME2VXMWyri2Z4W+jOIPC5k+Hb+vcgdwMAc7OhRA==
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfAQTA+tly+X+WcCw0v5LFJfk7t27wWKi8/fMfX7lhf3HsC0mPzo6wGISq6sag/k07J/kf/zQ6ScJErN/7fn3Sd3fGPuTu1dtAroJrE5w84+Vr98K2L+5qenbc+YMQ4+KC8PTF88YDJavZnCv/MAwJa6YwVXpDEPcVlGGgLZfDM657gzL5s5h0EptYtDjE2QojtJjeFvozjB3qy6Dm7MsyN0ANYOh2A==
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTs9wUl3Xet+P0fCByYRPS91Hdp2H9dlXv61ykzoPGsDFBg/wPGYmB2EL4X6TVB8BFMDKyXLZf7ZwHDSvsvUVySu3fvBouJzt8z9/mVF/Yfw7aY/OjoAItJrK5qDObTsH+S//FDp58kSMz+teffJ3V/Y+xP7l61CegmsDrBzT9Wvn4rYP/mpqZvz5kzDEeObGB4+uIZg8Hy1QyJFyoYpsQVM7gqnWGI2yrKwHmamcE5152hd64ug0GwA4MenyBDcZQew9tCd4a5W3UZ3JxlQe4GANytncw=
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTsr26rOVleXvYfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfIQkxujAlsv9s4Bhpf2XKC7J3bt3g80Tnb9n7vMrL+w/hm0x+dHRARaTWF3VGMynYf8k/+OHTj9JkJj9a8+/T+r+xtif3L1qE9BNYHWCm3+sfP1WwP7NTU3fnjNnGI4c2cDw9MUzBoPlqxkSL1QwTIkrZnBVOsMQt1WUQbJej8E5151h2dw5DAbBDgwsraUMxVF6DG8L3RnmbtVlcHOWBbkbAHoAnuA=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDkwMs7hdGsTsL7/celfZ2MiB5Za78DcDPftLc0+4Cykp2n/iclVl6Oiwv+d5KX+RkpL988fmNn7v3v0HAgcmEX0v9V0a9l9X5Z7+dcoMaDyr/Yd5j/QFp1ja/2CAAWYH4XuRXhMEH8HEwHrZcrl/FjCstP8SxSW5e/dusJjo/D1zn195Yf8xbIvJj44OsJjE6qrGYD4N+yf5Hz90+kmCxOxfe/59Uvc3xv7k7lWbgG4CqxPc/GPl67cC9m9uavr2nDnD0KPiwvD0xTMGg+WrGRIvVDAs2TuNwVXpDMMGttkMAW2/GJxz3RmWzZ3DoJXaxKDHJ8hQHKXH8LbQnWHuVl0GN2dZkLsBRzCfqg==
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/t3Aqeu7zlz1v6J3/qZOWfOOLDv7q9YeDXH/vLLrXeVjY0cWG65C38z0LN/yOXjUXLmjP0nLldVho4O+3uel/IXKSnZP39sbuP37t1/IHBgEtH3Ut+lYf91Ve7pX6fMgMaz2n+Y90hfcIql/Q8GGGB2EL4X6TVB8BFMDKTX/m3SZ6tV857bf9ge2f2jowNsnuj8PXOfX3lhf/Nx+cvImTPAYhKrqxqD+TTsn+R//NDpJwnW+9rz75O6vzH2J3ev2gR0E1id4OYfK1+/FbB/c1PTt+fMGYYeFReGpy+eMRgsX82wK9WEYUpcMYOr0hmGuK2iDAFtvxicc90Zls2dw6CV2sSQrGXAIGazluFtoTvD3K26DG7OsiB3AwAL/6YU
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTsnz82t/F79+4/EDgwieh7qe/SsP+6Kvf0r1NmQONZ7T/Me6QvOMXS/tubti89Z84CxZgdhO9Fek0QfGT/A+IGkF77t0mfrVbNe27/YXtk94+ODrB54iGfy3bLS9jffFz+MnLmDLCYxOqqxmA+DfvHqccncfpJgvW+9vz7pO5vjP3J3as2Ad0EVif4Qvj+Ve4N9m9uavr2nDnDcOTIBoanL54xGCxfzZB4oYJhSlwxg6vSGYa4raIMAW2/GJxz3RnOuU5n0EptYkjWMmDw4d3C0DKBjWHuVl2GZwp9IHcDAB5AqJk=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9rXnXy+o0VQONZ7T/Me6QvOMXS/gcDDDA7CN+L9Jog+AhJjNGBLZf7ZwHDSvsvUVySu3fvBpsncuqBilV0if3HsC0mPzo6wGISq6sag/k07J/kf/zQ6ScJErP/9HLKfq63l+xP7l61CegmsDrBzT9Wvn4rYP/mpqZvz5kzDD0qLgxPXzxjMFi+msG98gPDlLhiBlelMwxxW0UZAtp+MXhcWcewbO4cBoNgBwaW1lIGXc/VDG8L3Rkkq48yuDnLgtwNAC25ncs=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2+4KS7rtW/P4PBA5MIvpe6rs07L+uyj3965QZ0HhWBiiw/wFjMTA7CN+L9Jog+AhJjMmBLZf7ZwHDSvsPp39mbhQUAAk6iM7fM/f5lRf2H8O2mPzo6ADbIbG6qjGYT8P+Sf7HD51+kiAx+9eef5/U/Y2xPxgSMunxob9gdYKbf6x8/VbA/s1NTd+eM2cYelRcGJ6+eMZgsHw1Q+KFCoYpccUMrkpnGOK2ijJwnmZmcM51Z1g2dw6DVmoTg5hKL8O2umiGt4XuDKxrzjG4OcuC3A0Ae1CW+g==
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTsnz82t/F79+4/EDgwieh7qe/SsP+6Kvf0r1NmQONZ7T/Me6QvOMXS/gcDDDA7CN+L9Jog+AgmBtbLlsv9s4Bhpf2XKC7J3bt3g8VE5++Z+/zKC/uPYVtMfnR0gMUkVlc1BvNp2D/J//ih008SJGb/2vPvk7q/MfYnd6/aBHQTWJ3g5h8rX78VsH9zU9O358wZhiNHNjA8ffGMwWD5aobECxUMU+KKGVyVzjDEbRVlCGj7xeCc686wbO4cBq3UJgY9PkGG4ig9hreF7gxzt+oyuDnLgtwNAF0no7U=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfAQTA+tly+X+WcCw0v5LFJfk7t27wWKi8/fMfX7lhf3HsC0mPzo6wGISq6sag/k07J/kf/zQ6ScJErN/7fn3Sd3fGPuTu1dtAroJrE5w84+Vr98K2L+5qenbc+YMQ4+KC8PTF88YDJavZnCv/MAwJa6YwVXpDEPcVlGGgLZfDM657gy9c3UZDIIdGPT4BBmKo/QY3ha6M8zdqsvg5iwLcjcAHkWhAg==
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuz7G+RTdjPaX3659a6ysZEDyy134W8GevYPuXw8Ss6csf/E5arK0NFhf8/zUv4iJSX754/NbfzevfsPBPZfNta7iqxzsP+6Kvf0r1NmQONZ7T/Me6QvOMXS/gcDDDA7CN+L9Jog+AgmBtb7Numz1ap5z+0/bI/s/tHRARJzEA/5XLZbXsL+5uPyl5EzZ4DFJFZXNQbzadg/Tj0+idNPEqz3teffJ3V/Y+xP7l61CegmsDrBF8L3r3JvsH9zU9O358wZhh4VF4anL54xGCxfzeBe+YHB1aWMwVXpDEPcVlGGgLZfDKKXjjAsmzuHQSu1iSFZy4DBh3cLQ8sENoa5W3UZnin0gdwNAO+to2I=
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTs9wUl3Xet+P0fCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfIQkxujAlsv9s4Bhpf2XKC7J3bt3g80Tnb9n7vMrL+w/hm0x+dHRARaTWF3VGMynYf8k/+OHTj9JkJj9a8+/T+r+xtif3L1qE9BNYHWCm3+sfP1WwP7NTU3fnjNnGI4c2cDw9MUzBoPlqxkSL1QwTIkrZnBVOsMQt1WUgfM0M4NzrjvDsrlzGAyCHRhYWksZiqP0GN4WujPM3arL4OYsC3I3AFMjnsY=
eNqzf6n4bFngzJn293cueNHH+8T+0aztNpwzZ9nvnAkCs+yPBIj80ejocGDpKlY1j1Oz/8AABvZvt0Qfn15eYf/Eb/3MnDNnHNh391csvJpjf/nl1rvKxkYOLLfchb8Z6Nk/5PLxKDlzxv4Tl6sqQ0eH/T3PS/mLlJTsnz82t/F79+4/EDgwieh7qe/SsP+6Kvf0r1NmQONZGaDA/geMxcDsIHwv0muC4COYGFgvWy73zwKGlfZforgkd+/eDRYTnb9n7vMrL+w/hm0x+dHRARaTWF3VGMynYf8k/+OHTj9JkJj9a8+/T+r+xtif3L1qE9BNYHWCm3+sfP1WwP7NTU3fnjNnGI4c2cDw9MUzBoPlqxkSL1QwTIkrZnBVOsMQt1WUIaDtF4NzrjtD71xdBoNgBwY9PkGG4ig9hreF7gxzt+oyuDnLgtwNAFbWnyM=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0ODDNONMYeaTH/gMDGNi/3RJ9fHp5hf0Tv/Uzc86ccWDf3yCfspvR/vLLrXeVjY0cWG65C38z0LN/yOXjUXLmjP0nLldVho4O+3uel/IXKSnZP39sbuP37t1/IHBgEtH3Ut+lYf91Ve7pX6fMgMaz2n+Y90hfcIql/Q8GGGByEL4X6TVB8BFMDKyXLZf7ZwHDSvsvUVySu3fvBouJzt8z9/mVF/Yfw7aY/OjoAItJrK5qDObTsH+S//FDp58kSMz+teffJ3V/Y+xP7l61CegmsDrBzT9Wvn4rYP/mpqZvz5kzDD0qLgxPXzxjmCd0mMG98gODq0sZg6vSGYa4raIMAW2/GJxz3RmWzZ3DoMhawKDHJ8hQHKXH8LbQnWHuVl0GN2dZoDOZAdy5oXA=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdBOMaBFfMj4GJgfTav036bLVq3nP7D9sju390dIDNE52/Z+7zKy/sbz4ufxk5cwZYTGJ1VWMwn4b949Tjkzj9JMF6X3v+fVL3N8b+5O5Vm4BuAqsTfCF8/yr3Bvs3NzV9e86cYehRcWF4+uIZg8Hy1QzulR8YpsQVM7gqnWGI2yrKEND2i8E5152hd64uA0PCfoZkLQOGwCdHGVomsDHM3arL8EyhD+RuAFAupFE=
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuz7G+RTdjPa3wgp+xh7944Dyy134W8GevYPuXw8Ss6csf/E5arK0NFhf8/zUv4iJSX754/NbfzevfsPBA5MIvpe6rs07L+uyj3965QZ0HhW+w/zHukLTrG0/wGxD6xOMK5BcMX8GGQx+7dJn61WzXtu/2F7ZPePjg6wOtH5e+Y+v/LC/ubj8peRM2eAxSRWVzUG82nYP049PonTTxKs97Xn3yd1f2PsT+5etQnoJogdL4TvX+XeYP/mpqZvz5kzDD0qLgxPXzxjMFi+msG98gODq0sZw8pT9xnitooyBLT9YnDOdWdYNncOw5QVSQzJWgYMgU+OMrRMYGOYu1WX4ZlCH9CZzAANwam3
eNqzf6n4bFngzJn2b0rO3tmpKWL/aNZ2G86Zs+x3zgSBWfZHAkT+aHR0OLB0Fauax6nZf2AAA/u3W6KPTy+vsH/it35mzpkzDuy7+ysWXs2xv/xy611lYyMHllvuwt8M9Owfcvl4lJw5Y/+Jy1WVoaPD/p7npfxFSkr2zx+b2/i9e/cfCByYRPS91Hdp2H9dlXv61ykzoPGs9h/mPdIXnGJp/4MBBpgdhO9Fek0QfAQTA+tly+X+WcCw0v5LFJfk7t27wWKi8/fMfX7lhf3HsC0mPzo6wGISq6sag/k07J/kf/zQ6ScJErN/7fn3Sd3fGPuTu1dtAroJrE5w84+Vr98K2L+5qenbc+YMQ4+KC8PTF88YDJavZnCv/MAwJa6YwVXpDEPcVlGGgLZfDM657gy9c3UZDIIdGPT4BBmKo/QY3ha6M8zdqsvg5iwLcjcAHkWhAg==

[envertigo]

getting there... (nice progression on the graph) (and that car is a mutant that showed up for the screenshot)


I am saving everything from now on! lol

[honnza]

I can see you're running the applet in a standalone viewer. From what I remember, the cars appear to behave differently depending on whether they are running inside a browser. You may want to transfer your simulation there .

[envertigo]

Yikes.
Thanks for the warning.

I am running a normal sim, I am not inputting and receiving values from externalUpdate (didn't work, didn't have time to mess with it ) where I can imagine there will be browser differences.

Where I can't rule it out, I would imagine (but not confirmed) that ie, ff, chrome, stand-alone, embedded-ocx in c# app (I do that too) would use the same underlying control for the flash-app.

I can see how different versions (of flash) could behave differently though. I noted on a very slow box that a friend uses (like a single core amd 2200xp or something) that I see the drawing go a bit strange (wheels slipping into the floor and whatnot but otherwise performing normally (unlike the bug exploits I have seen).

With all that said, it is entirely possible that every configuration will perform differently. I would be curious to know if the simulation is performed on a reference before drawing the result (which may level the field) or if the scene as drawn by flash plays a part in the simulation.

(aside: I am getting to the point where I want to start messing with my own algorithm, but that would need a version of the sim that takes an input (cars), runs them, returns a result (scores/times), then waits for more input. I will digress on this point and post it somewhere else when I can articulate what I want better lol)

Bottom line for me is I have to run it that way to allow for work (like workwork) to continue while the sim runs... I think as long as they can be pasted into any same version of boxcar2d no matter the host, the sim is valid (imo)

[honnza]

The simulation is indeed independent on which browser you use. All that matters is PC/MAC and (AFAIK) browser/standalone. The wheels sinking happens all the time and this is what enables the top-scoring cars in Tight Squeeze reach their score. The time is measured assuming 30FPS, so slower machines will observe the clock ticking slower.

Concerning your other request, this is probably what you're looking for:
viewtopic.php?f=4&t=99

[envertigo]

Thanks for the link to the contest. That is the thing that I attempted to get working and could not due to time -- I do get the impression though that it works, just not as I expected. I will bring it up in that topic though, when I ask Lotharsson for all of his code and (lol) Ryan for his minus any personal IP

I think (again, might be wrong) that there is a difference between a sinking wheel that allows for wondrous things and a less-accurate draw on a slow machine that otherwise performs normally.

Still running (here and at home(numbers unknown on that one)):
(Generation: 425 Max Score: 328)