If there were a random specific required sequence then you'd have a 1 in 6 chance of getting the first pull right, 1 in 5 on the second, 1 in 4 on the third, et cetera.
1/6 * 1/5 * 1/4 * 1/3 * 1/2 = 1 in 520 chance of success
That's significantly worse odds than what we have actually seen. People have also reported collapse on the last wheel (though Cannith boots can potentially get you over the gap in that case).
Unfortunately, there are just too many possibilities to even begin to guess. The sequence could even be completely irrelevant. For all we know, every time ANY gear is pulled the game does something like 1d3; 1=Nothing, 2=Lightning, 3=Collapse.
Maybe disarming traps helps. Maybe two 'lightning' results, or a 'lightning' result before two successes, causes a collapse. Maybe the results are different on Tuesdays. We just don't know... and the amount of testing which would be required to evaluate all the possibilities is beyond insane.