Because the movement allows you to move the most weight. There is more to it but that's all you need to know.
I tried dips in both ways: vertical and leaning forward. Both led to pain in the shoulder joints, so I stopped doing them.
You will recall I mentioned earlier in another thread that I benched 285 for one rep in a gym competition in 1992-3, when I weighed about 145. I was probably at my strongest then. At that time, I normally did weighted dips with 75 pounds on the belt for 10-12 reps. (I think it might have been a bit more weight on the first set, but I can't get the picture of three 25-pound plates on the belt out of my mind, so let's go with that.) While the total weight is less than my one-rep bench, it was pretty much in line with my bench for reps. In fact, I started my bench press with a higher weight than 220, but for about half the number of reps. Fast forward to the late '90s, and I was no longer doing bench with free weights; I was using the Smith machine because of my shoulders. At that time, I recall wanting to up my dip weight and went to just over 100 pounds of added weight on my first set for a very brief period (a couple of weeks?). 105 comes to mind, but I can't be sure. And it was only for about 5 reps, before I started lowering the weight for subsequent sets and upping the reps a bit. That was still more than I could bench on the Smith machine for reps. It was in 2012 that I seriously started revamping my workout routine, because it was too much volume and, possibly, weight for me at that time. But I distinctly recall being able to characteristically do 14 reps with 70 pounds on the belt, a 45 and a 25. The reps were full range, but they were fast. I couldn't finish them quickly enough. My weight was in the high 130s then. It was a different gym, and I was no longer using a Smith machine for bench press, but a Nautilus-type machine. I was dipping more than I was "benching," and I was dipping with free weights. Perhaps it's because of the nature of my compromised shoulders, but at least in my case, I don't think you can make that generalization. As an aside, since 2012 and up until a number of months ago when I started doing very slow reps with only body weight, I was using a 65-pound dumbbell, which I held between my legs just above the knees, for 8-9 slower reps than when I was previously racing through them, but not as slow as I presently do them unweighted (obviously). I got tired of carrying the belt around with me in the gym (along with my own handle for pulley rows because the gym didn't have the kind I liked). That was when I decided to go full Thoreau.
Stated differently, and perhaps a bit more succinctly, are you sure that argument holds when you add your body weight to the added weight for a dip? Also, I would think that the dip is as natural a push movement as any other movement, including a horizontal push such as the bench press. In any event, I do both; I just prefer chest dips presently, and therefore do them first.
Interesting, but I'm not sure I'd place too much confidence on a sample size of one. Also, I don't understand why an exercise at a lower weight would get a lower reading unless the rep number was arbitrary and not to full concentric failure. Finally, I've read that the normalization and interpretation of EMG signals is a complicated area, and one that even researchers get wrong: https://www.strongerbyscience.com/emg-amplitude-tell-us-muscle-hypertrophy/ https://bretcontreras.com/can-motor-unit-recruitment-be-inferred-from-emg-amplitude/ Regardless, this stuff is above my pay grade.
A muscle fiber is either activated or it is not. It's zero or 100%. As to lower weights, that depends on the rep range, as explained by Doug McGuff in his book, Body By Science: “If you use a weight that is too light….you will recruit the slow-twitch fibers into service, but because they fatigue so slowly, by the time you have started to recruit the intermediate fibers, some of the same slow-twitch motor units will have started to recover. They will then recycle back into the contraction process, thus preventing you from ever engaging the higher-order muscle fibers.” According to McGuff, there is a similar problem with a weight that is too heavy, allowing for only one or two reps: All motor units (slow and fast) are activated, but the fast-twitch units fatigue so fast that “the set will terminate before you’ve had the opportunity to thoroughly involve and stimulate the bulk of your slow- and intermediate-twitch fibers.” The Goldilocks rep range seems to be 6 to 20. https://www.cbass.com/Carpinelli.htm Again, above my pay grade. I'm just passing along what I read.