A lesser-known aspect of extractor tuning is deflection, measured by the amount of the tensioning wall directly behind the claw that protrudes into the breech area and determined by the front pad that locates it relative to the breechface centerline.
Recently, I've seen several that have had excessive deflection, which can mimic the 3-Point Jam and the "Ka-Chunk" sound that goes with it...and can also cause the sometimes maddening intermittent failure to go to battery.
Excessive deflection has caused the ruination of many frame ramps by sending the amateur tuner to get Ye Olde Dremel for a feed ramp polishing, thus turning a simple problem into a complicated, expensive problem in quick time.
In the picture below is an example of way yonder too much deflection. In this instance...more than twice the amount that it needs. Depending on the distance between the two small, rectangular guide blocks... .480-.482, I like to see .010-.012 with a maximum of .015 inch. If all is within spec respective to that dimension, and the extractor and its channel, the bottom of the wall shouldn't need modification beyond lightly breaking the corner.
Study the photograph and see how the wall can easily act as a sort of "speed bump" to the case rim as it enters the extractor.
The owner of this one had tried to bevel and radius the bottom corner in an attempt to get the gun to behave, to no avail.
The distance between the guides was to spec at .482 inch, and the claw to wall depth optimum at .036 inch...and the wall showing in the breech was .030 by best eyeball estimate. Thankfully for him, he brought it to me before grabbing the Dremel. By carefully removing material from the wall and shortening the claw by a like amount, the pistol was up and running in about 15 minutes.
The basic problem was that the locating pad was too small, placing the wall and claw too close to the breechface centerline.