Well, it's quite late, I've been working hard all day & drinking a bit this evening
and those figures are making my head spin!
I'll try to have a look in the morning.
In the meantime, here's my (much simpler I think) way of working out the probabilities (which may be seriously flawed, as Zoozie said):-
Lets take
DogFather, with the reel layouts I posted above:-
There are 10 ways of hitting three scatters, the first is on reels 1, 2 & 3.
There is 1 scatter per reel, therefore there are 3 reel positions which will have a scatter showing.
So the number of ways you can hit those scatters is 3 x 3 x 3 = 27 ways.
The total number of different reel positions is 35 x 35 x 35 = 42,875
So the chances of getting 3 scatters on the first 3 reels is 42,875/27 = 1 in 1,587.96
(Or 6.30 times in 10,000 spins - you'll see why I quote that later...)
Does everyone agree with that? If not, what's wrong?
Now let's look at reels 1, 2 & 5:-
Reel 5 has 4 scatters, so there are 12 reel positions to have 1 scatter showing.
So the number of ways you can hit those scatters is 3 x 3 x 12 = 108 ways.
The total number of different reel positions is 35 x 35 x 40 = 49,000
So the chances of getting 3 scatters on these 3 reels is 49,000/108 = 1 in 453.17
(Or 22.04 times in 10,000 spins)
So I used the same calculation for all 10 reel possibilities, getting the number of times each combo should appear in 10,000 spins. Then just add them all together and divide into 10,000.
I originally did these calculations using fractions - and got exactly the same result.
Don't forget, my figure is for 3
or more scatters across the 5 reels, so also includes getting 4 or 5.
My spreadsheet:-