And you have yet to provide any information on what percentage of the total measurements your observed errors are. One, or even a hundred bad levels do not constitute proof that levels don't work.
You submit that the existence of observed errors makes the entire methodology flawed. By that premise the existence of an unreliable level would dictate that none of them can be trusted and people should stop using them, regardless of how many times the were used without any observed error.
I’m not telling you whether or not to trust them, but I don’t mainly due to contradictions with the Bible and secondarily all of recorded human history. Science is famous for the consensus / group think that defies all known logic and common sense. Here’s the intro to the article I’ve been quoting at creation.com/age-of-the-earth:
No scientific method can prove the age of the earth and the universe, and that includes the ones we have listed here. Although age indicators are called “clocks” they aren’t, because all ages result from calculations that necessarily involve making assumptions about the past. Always the starting time of the “clock” has to be assumed as well as the way in which the speed of the clock has varied over time. Further, it has to be assumed that the clock was never disturbed.
There is no independent natural clock against which those assumptions can be tested. For example, the amount of cratering on the moon, based on currently observed cratering rates, would suggest that the moon is quite old. However, to draw this conclusion we have to assume that the rate of cratering has been the same in the past as it is now. And there are now good reasons for thinking that it might have been quite intense in the past, in which case the craters do not indicate an old age at all.
Ages of millions of years are all calculated by assuming the rates of change of processes in the past were the same as we observe today—called the principle of uniformitarianism. If the age calculated from such assumptions disagrees with what they think the age should be, they conclude that their assumptions did not apply in this case, and adjust them accordingly. If the calculated result gives an acceptable age, the investigators publish it.