Slot Machine Wheel weighting estimating

How does one go about estimate the virtual reel weightings of a slot machine by observing the outcomes of the game?

Also large of a sample size does one need to closely approximate the outcomes?

Could you give me an example please.

Thanks,

Jim

/games/slots/appendix/1/

what I am referring to is this article, where mike estimates the reel weightings of a double strike machine by playing it roughly 4000 times, but i would like more clarification of how he comes up with the count on each symbol coming out to the estimated reel weighting of each symbol.

Quote: Jim123

/games/slots/appendix/1/

what I am referring to is this article, where mike estimates the reel weightings of a double strike machine by playing it roughly 4000 times, but i would like more clarification of how he comes up with the count on each symbol coming out to the estimated reel weighting of each symbol.

I have a feeling the assumption of 128 stops per reel is incorrect; it looks like he is actually assuming 72.

Assuming that there are 72 stops per reel - and there is no way to tell exactly how many stops a reel has, other than looking at the slot machine's (very proprietary, and very closely guarded, from what I am led to believe - except that a "multi-line" machine "should" have one stop per symbol) information sheets - divide the number of times a particular symbol appears by the number of spins (in this case, 3976), and then multiply by the assumed number of stops on the reel (in this case, 72).

I think your right he mistakenly put 128 stops when his virtual reel only has 72!

Thanks for the clarification.

Jim

Speaking of this, do machines have the same virtual stops on all 3 reels, assuming the reels are all identical? Like a double diamond, blazing sevens, etc. Or can each reel be weighted differently?

Can the JP symbol have a 2.5% chance on first feel, 1.5% on second, and 2% on third reels? Or do they all have to be the same likelihood?

Of course this can't be true for all machines, since not all machines have the same symbols on all reels. Stuff like 3x-4x-5x doesn't have 5x on first or second reel, 4x on first reel, etc. For the life of me I can't think of the name of another machine I'm thinking of, but there's a blue' sh symbol in third reel only and when you hit it, you go into a bonus round where you have 3 or 4 chances at picking. There's values above ranging from like $10 to $1,000, and you can either "take offer" or "try again".

Quote: RS

Speaking of this, do machines have the same virtual stops on all 3 reels, assuming the reels are all identical? Like a double diamond, blazing sevens, etc. Or can each reel be weighted differently?

Can the JP symbol have a 2.5% chance on first feel, 1.5% on second, and 2% on third reels? Or do they all have to be the same likelihood?

Of course this can't be true for all machines, since not all machines have the same symbols on all reels. Stuff like 3x-4x-5x doesn't have 5x on first or second reel, 4x on first reel, etc. For the life of me I can't think of the name of another machine I'm thinking of, but there's a blue' sh symbol in third reel only and when you hit it, you go into a bonus round where you have 3 or 4 chances at picking. There's values above ranging from like $10 to $1,000, and you can either "take offer" or "try again".

All the reels can be different.

It sounds like you're thinking of Top Dollar.

So in the example of 72 stop three reel machine, how many observations does one need to get a representative sample size?

Quote: Jim123

So in the example of 72 stop three reel machine, how many observations does one need to get a representative sample size?

You're getting into "confidence intervals" and "margins of error" territory here. Those are a little above my knowledge of statistics.

Also keep in mind that, on an actual machine, the total number of stops on each reel can be different.