Does a level get weaker each time it is retested? No.

Does a level get weaker the more times price comes back and tests it? This one had a false start – our first pass said yes, but that answer turned out to be wrong, and we're correcting it here in the open.

No, a level does not get weaker with repeat testing. Once you only count touches where price actually tested the level (rather than lumping in cases where price simply blew through without a real test), the hold rate does not fall. For the grid family it goes 67.1% at the first touch, then 72.6%, then 72.5% – it rises and then holds flat. For support and resistance it goes 61.9%, then 72.2%, then 72.6% – same shape. If anything, a level defends itself a little better on the second and later touches than on the first.

The mistake, named plainly: the first version of this study counted every break in the denominator, including ones where price never really tested the level at all – it just gapped through. Those untested gap-throughs are rare at the first touch (about 1% of grid breaks) but common from the second touch onward (77-79%). Counting them made it look like levels were "decaying" from about two-thirds down to about one-third holding, when in fact that drop was coming almost entirely from a change in what was being counted, not from a change in how the market behaves. That earlier, uncorrected reading should not be cited anywhere.

One thing that does not change either way: how the price move afterward resolves – whether it reaches its target once it does escape -- stays in a tight 52%-55% band regardless of touch number, both before and after the fix.