The problem isn't that people won't be happy where they end up. It's that it's possible (or even likely) that programs could end up interviewing lots of people needlessly. Imagine a middle of the road program. They have 10 spots. They get 300 applications - a mix of top performers, middle of the road candidates, and weaker applicants. The program is planning on inviting 80 people to interview -- an 8:1 ratio. It's possible that the program could do so and end up at the bottom of their rank list, or in SOAP. Trust me, this has happened at some point to every program (ignoring top programs). If 100 of the 300 applicants are a top performer, I can't invite them all -- I can only invite ~20 of them. If I filled my entire 80 spots with them, I'd probably be unfilled because they will have many other choices. My program is probably lower on their interest list. So how do I pick which 20 of the 100 to invite? That's the challenge here.
Whether this signaling will help or not is anyone's guess. If none of those 100 top candidates signals, what would this program do? Invite none of them? Pick 20 anyway?
Increasing interviews would be another solution -- but the time/money/resources are limited, there's only so many interviews we can do.
The key problem here is that programs receive more qualified candidate applications than they can handle. Ultimately we make decisions based upon really small issues -- has the applicant ever been in our area (which may not matter at all), does their prior research mesh with what we can offer (they may be uninterested in continuing the same research), reading letters of recommendation for small differences (this LOR says the student is "excellent", perhaps that means they're not as good because other letters from other writers say "outstanding"), etc.
Regarding the number of signals, it was kept low to avoid it turning into an app cap. Give applicants 15 signals and you'll never get an interview from programs you have not signaled. With only 5, (most) programs won't be able to interview only those whom have signaled.
Any change to the application process will help some and hurt others. Theoretically, signaling helps those applicants who have some clear idea when applying which programs they might be most interested in. Applicants whom are truly undecided would probably be best picking some programs to signal anyway.