The point of a stability test is that it tells you whether your system is probably stable, with the percentage chance of instability decreasing the higher the load that the test places on it and the longer you run the test for without encountering error. There is no "fail" switch that is magically tripped after X hours of testing, and there is probably no completely 100% stable system using ordinary desktop components (ie. non-ECC, etc). That being said, unless the issue is heat-related, there is a chance that whatever error is shown up by the very stressful stability test will show up during "normal" operation at even very low loads. Sure, the chance is reduced, but it's not eliminated. Whether you're ok with that probability is up to you - and if you're not running critical systems then sure, take the risk.

But don't kid yourself that your system is "stable". You've just proven that it's not.