Indeed, when we calculate the trade probabilities, we use overlapping trades, which are all the trades that met the trade condition in the past. There were 22 such days and the stop was hit 5 times out of 22.
However, we also need to judge whether the strategy worked well in the past when buying/selling during the analysis period. This is why we use non-overlapping trade in this case.
With non-overlapping trades we can compare the return generated by the strategy to a Buy/hold strategy during the past 120 days. We can see that during 51 days, the strategy produced 17.8%, while a B/H produced 26.47% during 120 days. Compared to the number of days invested the strategy is 1.59 times more efficient than a B/H strategy.
As far as the Risk/Return ratio, the expected profit per trade and the strategy efficiency are all in "green", then it makes sense to buy a pull-back to 108.17. I have an order waiting there.