I've cracked a few frames. All steel, but I only ride steel. Although people think of me as lanky, I'm a clydesdale, running weight between 180# and 220# over the years. I ride several thousand miles a year, but I don't rack up the serious miles of some other riders. Looking at other replies, I suspect weight is a key factor here. And the fact that I keep bikes forever.
My experiences-
'88 Bianchi Volpe cross bike. Cracked a right chainstay at the dropout around '00.
'00 Bianchi SL Cyclo frame (warranty replacement for above). Cracked a right chainstay at the dropout in '12, 2500miles into a TA tour.
'96 Gary Fisher Mamba mountain bike. Cracked a right chainstay at the dropout around '17.
'80s Peugeot PX-10 road bike. This was an ebay find around '08, which looked to have picked up some miles but likely had the original running gear. I rode it as my primary road bike for several years. Cracked a right seat stay at the top around '15. This bike is pretty spindly for a big guy, especially up around the seatstay where it cracked.
You may note the concentration of fractures on the right side. The chainstays would normally be in tension, but every pedal stroke some of that tension on the right side is relieved and may even become compression. Stress cycling is the enemy of almost any material and results in fatigue, with accompanying weakening.
Steel is supposedly more tolerant of fatigue than aluminum, as it has a fatigue limit. I.e. after frequent load cycles and weakening (talking on the order of a million cycles, plausible for a chainstay) steel settles in at a strength of about half of it's original strength. For materials encountering high load cycles, engineers just design to the fatigue limit strength.
Word is that aluminum does not have a fatigue limit. With frequent cycling the strength of aluminum eventually drops to zero. Yes, we make airplanes out of it, but the cycling is closely watched. For instance, one flight is one cycle for pressurization of the cabin. Short commuter flights rack up more of these cycles and the planes' maintenance and (life) has to reflect this.
Take heart, though, aluminum riders. Wish I could come up with a reference, but I saw a paper where someone destruction tested a bunch of frames, Al and steel, through like those million load cycles. The only frames that failed were steel. My recollection was that the writers seemed to be doing honest work and hadn't set out to prove something. Way back, aluminum was annoyingly stiff and they mostly relieved this with careful engineering. And any worthwhile engineer is at least thinking about fatigue limit during design, so I expect this gets at least some attention during design.
I have the good fortune of having a basement welding shop and except for the warranty replacement, all those frames have been repaired and are still in the stable. The Bianchi is a daily rider serving gravel, commuting, and light touring duty. The Peugeot is a fixie now (in my world the fixie is a novelty ride, so it gets easy miles). I brazed up the crack in the Gary Fisher dropout. I don't consider that a permanent repair and a replacement frame is in the jig.
