But it's a massively unreliable method for people who are not trained to translate
3D reality (4D if you like.) into a 2D picture.

i'd be interested to know if there are any kind of viable statistics showing the number of additional, unnecessary victims, resulting from this?

When you have a gun (that would be an actual Glock, not a comb-pretending-to-be-a-Glock) shoved against you and you are threatened to be shot on the spot, sometimes out of that kind of experience you actually do remember almost every detail.

I'm not a fairly good sketch artist but I am one of those "visual" people.  I gave the police a very detailed description of the man who robbed me at gunpoint.  I described his height, approximate weight and age, hair, and four specific articles of clothing -- ballcap, t-shirt, pants, and shoes.  I  gave them specific details about features on his face (eyes, nose, facial hair).  It helped (me, at least) that just before the actual robbery, I had gotten a steady look at the guy for 15-20 seconds.

A couple weeks after the incident, I was shown a photo array of eight, and within seconds (as in, almost immediately) I picked out the person I believed to be the gunman. All eight subjects in the pictures were within the age range of the description I have given police, and had the same hairstyle and facial hair characteristics as the man I had described; no piercings or tattoos on anyone in the photos, and no piercings or tattoos on the guy in my description.

Not long after I did the i.d. from the photo array, the police got a hold of a security camera shot of a guy swiping my stolen credit card at a gas station fewer than 6 hours after I had been robbed (and less than a half-mile from the site of the robbery). The clothing worn by the man on the security footage was exactly what I had initially described to the police, with exact colors on all the correct items (one article of clothing was dark blue, two more were black, the pants were baggy, dirty blue jeans cut off at mid-shin).  The insignia on the t-shirt of the guy in the footage was also an exact match to the description I'd given police (a hollow, oblong white diamond that almost went the width of the front of the tee).

Is eyewitness i.d. unreliable?  In the aggregate, yes.  But each case really comes down to the visual acuity and memory of the specific eyewitness.  In my case, I observed a lot and forgot nothing.  (Like I said, it helped that I got a good, long look at the guy beforehand, and that I have one of those memories.  If either of those factors not been the case, it could have been quite a different story.)

BTW, the defendant in my case was out on bail for more than two years before the case went to trial. Besides my testimony, we had him on security camera (visual clarity: exellent and close) swiping my credit card, and driving a car that was id'd as belonging to his mother.  The guy had no alibi. To this day, he swears it was his "brother" who did it.  (No record of his having a brother, wouldnt ya know, let alone one who looks enough like him to be his twin.) In other words, as legal procedure goes, this guy was as guilty as you can get.  And yet, even after this man had used a firearm to assault and rob an innocent, taxpaying citizen, and despite the fact that he had multiple drug possession counts on his prior record (junkie + firearm = public menace), the legal system allowed this guy to walk around on the streets for 28 months before bringing him to trial.  To paraphrase cpinva, "Anyone with a sense of justice [should be bothered by] the prospect of the actual perpetrator still loose on the streets, free to commit more crimes, because of shoddy work" by our so-called justice system.

Are innocent people behind bars a stinking offense to the notion of justice?  Absolutely.  So are violent, predatory criminals who are turned loose on the general population because the legal system is incapable of dealing with them (or anyone else) speedily (or incarcerating the predators long enough).  Which is a bigger problem in this country?

I have an outrageously good memory. Everyone who knows me well eventually figures this out and points it out to me (they often forget that they have and do so more than once ;) ). You don't know me, so you'll have to take my word for it (or not, if you prefer). I actually think that I am likely a bit unusual neurologically.

At any rate, I can recite trivial conversations I had 10 years ago word for word if they made an impression on me for some reason. People are often surprised when I see them after some years and ask after relations they mentioned in passing- it's too bad I have the wrong temperament for politics as I imagine that would be a useful skill for a meet-and-greeter.

Because I do have a very good memory, I notice the degree to which people begin to distort events very shortly after the fact. When I was a kid I didn't understand this, and I just thought that most people were terrible liars.

Beyond that, I have also become aware that even though my memory is much better than most people's, it is also unreliable. My subjective take is that I have a pretty accurate record of every waking moment since I was a small child stored in my head. The reality is that I don't.

On a couple of occasions I have been approached by people that I did not recognize who seemed to have a good idea of who I was. It turns out that I had met them and made an impression on them, but the reverse was not true- they had made no impression on me, and I had entirely forgotten them.

I read a memoir by the novelist D M Thomas about 15 years ago (if my memory is accurate, and it might not be so). It is titled, aptly, "Memories and Hallucinations". In it he describes an event that was "burned" into his memory- during the second world war a bomb fell in his front yard, and he describes the experience luxuriously, in terms of every sense except touch.