In my opinion, a cracked receiver is an indication of a manufacturing defect, either design or materials, at any round count. In order to crack, the receiver must flex repeatedly beyond it's plastic limits. It should not. The Sportsman 58s would develop cracks at high round counts and/or heavy loads, likewise the LW20 1100s (or so I was told by Wayne Leek, though I have never personally seen one of those). Remington redesigned both of those receivers to eliminate the problem(s).
Everyone knows the MIM extractors are frowned upon. The reasons for this are that they can break, and the reason for that is that the MIM process is much more sensitive to control conditions and is therefore much more likely to produce "bad" parts than forging, but there is no economical non-destructive way to test the finished parts to separate the good from the bad. MIM doesn't produce a lot of bad parts, just several times more than forgings. No good way to test the forgings either, but in that case the process is much less likely to produce a bad part. A good MIM part should work every bit as well as a forging - if properly designed for the application, etc. There can also be problems with parts produced by forgings; material fluctuation, temperature fluctuations, etc.
In the case of 870s, if it was a design defect, they would all fail at around a certain round count. Particularly target guns where the loads are similar. They don't. So I believe the differences are caused by variations in the material or manufacturing process. And while these do lower the failure point, even the "bad" ones seem to be good for 100,000 rounds plus.
I have broken a forged extractor on an 1100, and they are the same as an 870 extractor - only one - in over 50 years and hundreds of thousands of rounds, but they must have been being pushed by the semiauto design, because Remington made them larger when they introduced the 11-87.
I have seen one cracked receiver on an 870 where the receiver was actually worn a little thinner in the area of the action bars, at reportedly over 300,000 rounds, and the crack was in the same area. I have seen another one at the top front of the ejection port at a reported 150,000 round count. I would say the second one had some type of material or manufacturing process defect that occurred during manufacture.