Well you have a pretty complicated question there.
First off...can you tell me about the strap failure you experienced with the Triumph please. That's a pretty popular seat around here and the harness design is actually used on multiple seats in Evenflo's lineup, but this is the first time I've heard of a design flaw.
Now on to your question. The amount of variables between vehicles, crashes, the effect of crash forces on different parts of the body, and child sizes makes a direct comparison of a seat's crash performance to complex to put into a meaningful rating system.
To give you an example here:
lets pretend we invent a point system for different crash scenarios. For the sake of simplification, there are going to be only two categories
-ffing head injury criteria with a tether and ffing head injury without a tether without a tether- (HIC is a measure of likelihood of head injury, with or without head contact)
and the seats tested are only for children weighing 30-35# and 30-34" tall (a small enough window to ensure a pretty consistent size)
Now seat performance with a tether and seat performance without are both very important, but only about 40ish percent of people in the US actually use a tether (please pass the word about how important they are to all your friends) so we can weigh the results of the without tether as 60% of the final score.
Seat 1: HIC with a tether says there is a 20% chance of serious head injury so we will give it an 8 for tether score. Without a tether HIC says there is a 50% for so we will give it a 5.
weighting the scores as I said above, the combined score for the seat is 5.9
Seat 2: HIC tethered 2(80% chance of serious injury), and without tether 6 (40% chance of serious injury) leaving us a total score of 4.6
Seat 2 seems like the safer seat here, and if you are using it untethered, it is safer...but seat 1 is significantly safer than seat 2 when used without the tether.
If there were just 2 categories the scores could easily be split so people could make a choice based on what the expected usage would be, but there are actually a lot more variables.
Instead the way the system works is threshold conditions have been set for all these dozens of criteria of what performance is needed to reasonably expect a kid to survive a crash with only minor injuries. These thresholds are pretty comprehensive and are the result of years of research into what the tolerances of the human body are and what the likely effect of vehicle variables on the carseat are. A seat that can't meet one of the thresholds doesn't pass. There are literally dozens of ways a seat can fail. It's like taking a hundred question multiple choice test where you have to get 100 correct in order to pass the class. You have to know your stuff if you are going to make it.
What has come out of this system are extremely safe seats. We know certain features are showing promising results in categories like side impact safety that are not currently part of the testing procedure (in large part because it takes years of real world data to figure out what the threshold allowances need to be and in smaller part because seats are already so very safe that the additional expense has questionable potential returns )
and that certain features make correct use more likely (a HUGE deal when we are looking at a 90+% misuse rate) but the overall picture of safety is too complex to simplify into something the general public could use. For now the most important advice we can give is the "Choose a seat that fits your child well, fits your car well, and is easy enough for you to use properly every time" line.
If you do that, you will be making a very safe choice for your child, I promise.
