However do you think the show would have got commissioned if its first series had confirmed every myth they tested? Would it have got recommissioned?
Do you think they took that chance, or do you think they took a range of myths, some of them likely to be confirmed, some likely to be busted?
Do you think they plan the split between likely true/likely false to give a ratio that makes for entertaining (or more accurately profitable) TV?
If you start with broadly the right ratio, then a few surprise results is no big deal. Firstly, the "best" percentage is really a range (nobody is really going to see a huge difference between 65% busted and 70%) so you can afford to be wrong in your predictions without real harm. If you are wrong too often (so that you are busting 90% of the myths after half the filming), then change the mix of those myths you are still to film to include more likely to be confirmed. Mix over the series comes out reasonable and you simply show them in a different order to that they were filmed in.