The 23.6 was an exact retracement top and a chance to join in the downtrend.
You do a very good job of illustrating your point, but I think there is lots of hindsight bias going on here. If we are to take a short at the 23.6 like in your second chart, we have to also assume that a short should be taken on the first chart, and there is blows past 23.6. Sure, you can have a filter for when to enter, and at least this 23.6 provides a nice double top, but then if using double tops, then this doesn't work on the 50% and 61.8% like in your first example since price blows through those levels the second time it comes up.
I do think we need to see the whole chart to see that your first example is rather cherry picked. This example is my fib retracement C, which starts at point D. It looks almost random really. If we picked another swing high, like at point E, which produces the B fib retracement, then we see that price didn't really turn at the 23.6, and didn't make it to 38.2.
The first fib, A, did curiously also have a nice short entry at the 38.2, but it blew through the 23.6
So I think if we are to look at this in the context of using every major swing point as a possible beginning for our fib retracement, and then try and take shorts at the 23.6, the 38.2, the 50 and the 61.8, we might find that there are too many combinations. Once we start to mess around with different stops to try and maximize the wins, or try and use different reasons for entry, which means some good trades will be skipped, or some entered that fail, the magic of fib retracements might disappear.