Well, say the sabot does get stuck. Since it doesn't even seal that well (not well enough to handle pressure buildup, unlike piston seals), nor is it designed to handle pressure itself, all pressure buildup in the barrel would push out that spot. This does leave momentary pressure on the barrel, but the sabot will break out first before damage is done to the barrel material, or the air will escape through the sabot seal.
It's like over-pumping a Douchenator; the schrader valve is more likely to pop out first than the PVC to crack. Or maybe I'm wrong, as someone (whom I'll keep anonymous) who tried putting dry ice in a 2L bottle, sealing it up, and running away, found that the whole bottle exploded; the cap didn't shoot off first. Perhaps that was due to very fast pressure build-up though, but the threads on a 2L bottle are considerably thicker than the plastic so it may have been that.
For the 50mm, air will definately escape through the breech's seals if the sabot somehow got stuck and held pressure perfectly. However, both the barrel and the coupler are pressure rated, so if both points were to seal up perfectly (or well enough to hold some pressure for a while), the pressure in there would be pretty low, I would guess half of that which it was pumped up to.
The 76mm, on the other hand, is a tad bit more risky. I have cellcore pipe on it, and if you stuck a piston down the barrel, somehow held it in, and powered up the PC and fired, it would damage the pipe.
Lewis himself noted that the biggest safety concern is with excess water causing stress on the barrel. The barrel should be welded on and secured, while any extra water that comes from balloons exploding should be let out after shooting. My biggest fear is the schrader valve shooting off, but neither of us had have problems with that, and if it starts leaking, I'll just epoxy the threads and put putty over it.
So with that said, pressure buildup in the barrel is the least of my concerns. As Silence noted, some materials can take high stress for short time periods, and in order for a sabot to get stuck, there needs to be something very strong that keeps it from getting pushed out by 40-60 PSI. In typical water balloon launches, the barrel will experience stress for maybe 200ms. The fact that so much air can't even be felt going out of the breech seal assues me that all of it gets directed to the projectile just fine.
The video itself is running at a pretty low volume. Volume fluctuations on youtube are pretty common particuarly when simple video editors (like Windows Movie Maker) are used and provide limited control over the sound)
I have a few further modifications to work on before a water war this Thursday. I'm going to trim down the barrel as it's length is unfriendly to lower pressure shots (I may not have shown this in the video, but sometimes the sabot wouldn't even leave the barrel, while the balloon shot got considerably less range) and it's a bit cumbersome to carry around, though not as bad as the Douchenator which is now already half the height of what it used to be. I also have to get the bike pump orientated vertically to allow for proper pumping. 1-hand pumping is too difficult due to the force needed.