I didn't think pulling to set position would work harden the spring, in that position it is stressed where the arm starts. By work hardening i meant rapid polishing at that point, a bit like polishing a pivot work hardens as it compresses the steel. I doubt polishing by hand would achieve much in that respect though. Its finished, arm polished up mostly at the join to the bridge's main body. I'm ok with it, the screw holes aren't great as i had to open them up by redrilling and positioning it was difficult, I'm not much use with a loupe, opening up with a file might be a better option for me or i could just use the correct drill size 😅. And the detent is way too deep, i had to guess that with the stem release out of position and sat on top, but i only took one measure and went for it, no slowly slowly catch your monkey 😅. First go I'm happy , well sort of, it works and thats a big thing for me, next one will have a bit more finesse.  Anyone thats interested, after filing, i used a 2000 grade home made diamond  micro file and then 20 micron film, the film is much better than wet and dry, more stable to use and doesn't shed cheap grade grit everywhere , then auto polish on a sponge pad.

I'm assuming that every time you set the watch you are work hardening the detent spring, maximum hardening is  where it meets the plate due to maximum deflection.   That's why it snaps there.  The Young's modulus may be the same but after it's reached its maximum yeid strength it breaks.  My mechanic engineering is very rusty, correct me if I'm wrong. 

Ah ok yes that makes sense to polish it where the arm starts to form from the body of the bridge, i thought you meant the underneath of all the arm.

this is something I've never quite understood about the some of the Swiss companies. In 1957 Omega was using 9010 for the keyless parts with epilam. there's been a slow migration towards using heavier lubrication's but still typically oils and epilam to keep them in place. When it seems like 9504 works so much better.  

OK, welcome in the world of alarm clocks... I guess the 4th wheel is dished because it is from another movement. If it was not dishet, then it would not mesh with the pinion of the escape wheel, am I right? The marks of wear on the 4th wheel pinion doesn't corespond to the 3th wheel table position, at list this is what i see on the picts. Calculating the rate is easy - there is a formula - BR = T2 x T3 x T4 x T5 x 2 /(P3 x P4 x P5) where T2 - T5 are the counts of the teeth of the wheels tables, and P3 - P5 are the counts of the pinion leaves. Vibrating the balance is easy - grasp for the hairspring where it should stay in the regulator with tweasers, let the balance hang on the hairspring while the downside staff tip rests on glass surface. Then make the balance oscillate and use timer to measure the time for let say 50 oscillations, or count the oscillations for let say 30 seconds. You must do the free oscillations test to check the balance staff tips and the cone cup bearings for wear. This kind of staffs wear and need resharpening to restore the normal function of the balance.