Each time the OP went to pick the 6 hadn’t been picked. So at the time the OP went to pick those were the odds that he would pick the 6 no?
The chance of the OP being allocated a 6 in each of the draws is 1/17, 1/18 and 1/17 respectively
The chance of the OP getting a 6 in say the first draw *given that the previous 7 balls already drawn were not 6* is 1/10 as there are 10 balls left (at that point in time), one of which is 6.
When you make an observation you have additional information, if you've observed some prior draws and have seen that a 6 wasn't drawn then for each of those draws that didn't result in a 6 the chance of you drawing a 6 increases.
But what are you trying to represent when combining this for each draw? You're not being clear/haven't defined that and it's kinda muddled/doesn't make any sense.
Why would he, at any point in time, know of just the first 7 balls in all three draws at once? Save for some scenario where he owns a time machine and observes each of the draws up until he gets allocated a ball and already notices that in each case there is no 6 drawn?