I'd like to make a revision to one of my previous posts. I seem to have inaccurately measured a couple of gates.
14.7 - 15.4 (.7) --> 14.6 - 15.4 (.8)
17.2 - 19 (1.8) --> 17.7 - 18.6 (.9)
I don't know what happened there, but I guess it never hurts to double-check your measurements. In this particular case, the error contributed to my thinking, mistakenly, that the last number of the combination was 18. How embarrassing.
This leaves no gate that really jumps out at you, at least not based on gate width alone. They are all in the .7 - .9 range. Let's look at those gates again:
1.1 - 2
4.5 - 5.2
7.8 - 8.5
11.1 - 11.9
14.7 - 15.4
17.7 - 18.6
21 - 31.8
24.3 - 25.1
27.6 - 28.5
31 - 31.7
34.3 - 35.1
37.8 - 38.7
If we eliminate all those ranges that do not contain both an integer and at least .3 dial increments on each side of that integer (figuring the manufacturer would only use integers and would have those integers reasonably well centered in the gates), we are left with a much shorter list:
14.6 - 15.4
27.6 - 28.5
So let's look at how those two gates compare to others, looking for uneven placement of gates around the dial. It's easiest to group the gates by the ones digit:
Group A:
4.5 - 5.2
14.6 - 15.4 (true gate candidate A)
24.3 - 25.1
34.3 - 35.1
Group B:
7.8 - 8.5
17.7 - 18.6 (true gate candidate B)
27.6 - 28.5
37.8 - 38.7
Notice that candidate A is somewhat clockwise of where it seems to belong in the pattern. Candidate B, on the other hand, fits in pretty well with the pattern.
So candidate A it is! (Because of its non-uniform position in the pattern.) The third number of the combination, then, is 15. (g[3] = 15) And we go from there. The combination turned out to be 7-25-15. It would have been found rather quickly if it were not for my error.
[Note: I've found rather large true gates like what I *thought* I found here on Master "Sphero" padlocks.]
