Occam's razor - the simplest explanation is quite often the best explanation. In this case, the simplest explanation as to why the system "is inconsistent" with detection is that intentional feeding is inconsistent, or rather, the actions people can take that would be counted as intentional feeding are not cut and dry.
Your case of 24 deaths running it down mid is probably cut and dry, but this is an extreme outlier of an example. Remember, one single death fed intentionally counts as intentional feeding. How do we detect all of those? And how to we reduce false positives?
You're coming at this problem in an intelligent manner, which is a far cry from the ZOMG RITO NEVER PUNISHES INTERS shit that we have to deal with here on a daily basis. But when asked, I don't think any of us could come up with the programming necessary to detect intentional feeding in all cases, even in all egregious cases.
Take your example here. Thresh died 24 times, but his team is fed as fuck. Could he legitimately be intentionally feeding? Absolutely, and probably even likely. But since he has 12 assists and his team is fed as fuck, is there also a chance he was dying so his team could crush them in team fights over and over? Absolutely.
Sion, Zilean, and Garen's scores are comically bad compared to the winning team. Were they intentionally feeding?
