You are sort of comparing apples and oranges really but more importantly you are taking the worse anomaly after the event and using that as a statistic in your roulette example.
If I roll a dice 100 times and then check for the number that came in least I am just checking any random variance but if I first select a number and then roll the dice 100 times the chances of that number coming in an irregular number of times is 6 times less likely.
This is important because the Scatter and wild symbols are the most important symbols and if you were to check you wouldn't find the same anomalies occurring with other lesser symbols such as J,Q,K for example. (taking account of the relative probabilities)
For the probability of scatter landing (where 1 scatter is on the reel such as reel 1 and 2 on DOA)You need the amount of symbols on the reel divided by 3 for probability (assuming no weighting)
I doubt DOA reels are 50 symbols long, certainly look a lot shorter when lags and spins round full. If it were 30 symbols long the scatter should appear 1 in 10 spins on average.
Of course natural variance would dictate it would sometimes not hit in many more spins and sometimes come in much more frequently.
So over 150 spins for example if it came in twice over the first 100 spins (a very regular occurrence on DOA) then it would need to hit another 13 times in the next 50 spins to keep the average (like every 3 spins for 50 spins! Doesn't happen)
Obviously we would need a much bigger sample and the anomaly would not be quite so obvious but it certainly exists.
More telling is the fact that the Wild symbols also seem to come in much less regularly on reels 1 and 2 over the same periods.
Another anomaly seems to be that it is nigh on impossible to hit the bonus round when hitting the 1st two scatters.
I have triggered it once out of well over 50 attempts which is crazy but granted that could be just bad luck with such a sample size.
I expect the average for trigger free spins is somewhere around 150 spins or so according to other peoples stats yet I estimate I'm hitting every 250 spins if I'm lucky! And that is of a very decent sample size so it is massively irregular.
Maybe I will hit the bonus round every spin for 200 spins though and it will event out?
As for a wildline, if it 1 in 100 chance then yeah you could go a thousand rounds without hitting it but you neglect the flipside which you could also hit a wildline a few times in a hundred rounds.
Free spin rounds on most slots are less likely than a wildline in bonus rounds - around 1 in 150 rather than 1 in 100 and it you would be extremely unlucky to go a thousand spins without a trigger.
Thing is all these anomalies are ALWAYS in the slots favour.
I'm certain there is no game I have played over 100k spins and I'm ahead of the curve on.
Any game I hit well on I always have bad losing streaks on to equal things out - doesn't seem to work the other way around though which is a law of probability I am not familiar with.
I have lost thousands of dollars low rolling this year alone and very seldom have a chance of cashout.
I guess if your experience is different then it is hard to get your head around but I've been playing online for many years and my results seem to get worse and worse.
I'm currently trying to cut back on my playtime because even on the rare occasion I hit something and get ahead the casino (whichever) will just go completely dead all games - always. Yeah you can expect to lose if you keep playing but what is with the shut down and 65% RTP?
Anyway, I just want to hit a damn wild-line or 5 scatters on DOA just for the sake of it now.

Blind squirrels and all that.