I find this surprising. Specifically surprising. After the collision, I downloaded the last unclassified OES for both satellites prior to the collision and propagated them using SGP4, the model they are designed to work with. The model showed that they would pass within 70 meters of one another. Iridium is maneuverable, and supposedly maneuvers whenever there is a prediction that they will pass within 5 km (5000 meters) of another satellite. It is possible that they only check once a week or so and the OES came out within about 24 hours of event. The problem of collision avoidance is N-squared, the complexity goes as the square of the number of satellites, since I only had to check two satellites my problem was N-squared simpler. Still Iridium LLC only needs to worry about their 66 (maybe 65 now) satellites times a few thousand others.
Yes, someone paid me to do it.
Please do keep in mind that many analysts believe that the Iridium “incident” was a space war move. Take it for what it is worth, but a significant portion of analysts believe that this was a test to show how satellites could be knocked out of orbit WITHOUT an anti satellite missile launch
I can’t say any more at this point