Question Many Question about RTP/Variance

[dunover]

That would be an effective change to the reel layout The reel map is set up so that every possible combination of symbols occurs once, though that may mean that certain wins occur many times and others very few times due to the frequency of occurrence of each symbol on each reel. If you want to change the frequency that a specific combination occurs (3x9s for example) you by default have to change the reel map.

The second example is different and that's a change in paytable. If they're changing the paytable when you're changing coin size this should be apparent via quick inspection. I still don't really like it tbh, but it's an upfront change where changing the reel map cannot be easily observed by the player. I haven't seen that happening, but if they're doing it offline it wouldn't totally surprise me to find this happening online.

True, but not physically as the reels are simply a manifestation of the result pulled by the RNG. If the results pool was 'adjusted' the reels would simply be static; you'd basically see less 3x9's wins with the reels remaining untouched. That's the advantage of online slots as opposed to land-based mechanical ones. The scope for 'manipulation' is far greater.

 

[ThePOGG]

True, but not physically as the reels are simply a manifestation of the result pulled by the RNG. If the results pool was 'adjusted' the reels would simply be static; you'd basically see less 3x9's wins with the reels remaining untouched. That's the advantage of online slots as opposed to land-based mechanical ones. The scope for 'manipulation' is far greater.

We're talking semantic here. The RNG pool tied to specific symbols creates a physical map of the reels. One is the same as the other. If I want a reel that's 100 symbols long and I want a cherry symbol to occur 1 in 10 times I program that symbol to have a probability of 0.1. If I then change that probability to 0.01 (1 in 100), I've changed the reel map. Now only one of the 100 symbols is a cherry where previously 10 of them were. There's only two ways to accommodate this change - either I have to increase the frequency of other symbols or I have to reduce the length of the reels (or technically we could use a combination of the two). Either option changes the reel map.

Even where you start making certain stops on a physical reel more or less likely (weighting the bins), all you're doing creating an effective virtual reel that's longer than the physical reel. So for example if you had the following 10 bins on a reel with the probabilities in brackets:

1 - 10 (0.2)
2 - J (0.2)
3 - Q (0.15)
4 - K (0.12)
5 - A (0.09)
6 - Cherry (0.08)
7 - Orange (0.06)
8 - Lemon (0.05)
9 - Gold Coin (0.04)
10 - Star (0.01)

You have a "physical reel" that is 10 symbols long, but an "effective reel" with 20x10, 20xJ, 15xQ,12xK, 9xA, 8xCherry, 6xOrange, 5xLemon, 4xGold Coin, 1xStar. Any time you change the probability of any symbol occurring (by altering the probability of a specific bin) you alter the effective reel map. Bin probability is really just a shorthand way of describing a reel map.

Then of course you have to take into consideration the order the symbols occur in as this can impact concurrent wins on different lines, which again creates a different dynamic. The only point where order isn't a concern is where each symbol is a separate reel.....

 

[dunover]

We're talking semantic here. The RNG pool tied to specific symbols creates a physical map of the reels. One is the same as the other. If I want a reel that's 100 symbols long and I want a cherry symbol to occur 1 in 10 times I program that symbol to have a probability of 0.1. If I then change that probability to 0.01 (1 in 100), I've changed the reel map. Now only one of the 100 symbols is a cherry where previously 10 of them were. There's only two ways to accommodate this change - either I have to increase the frequency of other symbols or I have to reduce the length of the reels (or technically we could use a combination of the two). Either option changes the reel map.

Even where you start making certain stops on a physical reel more or less likely (weighting the bins), all you're doing creating an effective virtual reel that's longer than the physical reel. So for example if you had the following 10 bins on a reel with the probabilities in brackets:

1 - 10 (0.2)
2 - J (0.2)
3 - Q (0.15)
4 - K (0.12)
5 - A (0.09)
6 - Cherry (0.08)
7 - Orange (0.06)
8 - Lemon (0.05)
9 - Gold Coin (0.04)
10 - Star (0.01)

You have a "physical reel" that is 10 symbols long, but an "effective reel" with 20x10, 20xJ, 15xQ,12xK, 9xA, 8xCherry, 6xOrange, 5xLemon, 4xGold Coin, 1xStar. Any time you change the probability of any symbol occurring (by altering the probability of a specific bin) you alter the effective reel map. Bin probability is really just a shorthand way of describing a reel map.

Then of course you have to take into consideration the order the symbols occur in as this can impact concurrent wins on different lines, which again creates a different dynamic. The only point where order isn't a concern is where each symbol is a separate reel.....

My point was that video slots exist only as programs and yes, while originally the math was based on a reels set with defined possibilities people often make the mistake of 'seeing is believing' and assume the reels they see are always tied to the event as opposed to video graphics generated after a press has ordered a RNG result. So if I implemented one of your changes above to the result probabilities the RNG would pick from the amended pool but the reels/graphics would be exactly the same to the player, no need to change them! All you need to do is make a tweak to the math, not the reels. Of course if the player could get hold of the reel maps, run a couple of million spins through then maybe they would notice they are getting an unrealistic lack of 3x9's but we all know that won't happen!

 

[ThePOGG]

My point was that video slots exist only as programs and yes, while originally the math was based on a reels set with defined possibilities people often make the mistake of 'seeing is believing' and assume the reels they see are always tied to the event as opposed to video graphics generated after a press has ordered a RNG result. So if I implemented one of your changes above to the result probabilities the RNG would pick from the amended pool but the reels/graphics would be exactly the same to the player, no need to change them! All you need to do is make a tweak to the math, not the reels. Of course if the player could get hold of the reel maps, run a couple of million spins through then maybe they would notice they are getting an unrealistic lack of 3x9's but we all know that won't happen!

With regard to the graphics you see spinning past - yeah you're totally spot on, they're utter garbage that means nothing and is often programmed to show 'near misses' to make you feel that you were closer to a win than you are.

However, probabilities aren't the whole story even when it comes to videos slots, specifically because it's not usually a specific symbol your choosing from. Assume a 3 symbol high play area. You'll have a reel strip and what you're actually choosing from isn't a random selection of 3 individual symbols based on the probability of each symbol, but a single random set of 3 adjacent symbol on the reel. In other words only one selection is made to select all 3 symbols rather than 3 individual selections. So if we consider a reel with the 10 symbols ordered numerically the possible results are 123, 234, 345, 456, 567, 678, 789, 89T, 9T1, T12. There are only ten possible combinations. Rather than 1000 possible combinations that result if each symbol is chosen independently. This gets more complicated where the reels get longer (the whole purpose of virtual reels) and where symbols start to appear multiple times in different positions. Where position is a factor the reel map is always changed by making changes to the probabilities as this has to change the order symbols appear in.

Where there is no order to the symbols what we're really talking about is 3 separate reels within the reel with each symbol being chosen independently of every other symbol and at that point there's no difference between the reel map and the probabilities. There are a number of games on the market that function in this manner, but I believe most of them specifically define when this is the case. They can also be identified if you start finding the same symbol turning up in adjacent spaces with no mention of stacked symbols.

The point is that the reel map is - for most games - is still more than a hypothetical. Alongside the frequency of occurrence the order has to be defined. To change the payouts of a game you either have to change the paytable (ok as it's apparent to the player when this happens) or change the probabilities of occurrence of specific symbols which due to the order implications does change the reel map. This second option wouldn't be ok as the player would have no way of knowing that the game had been changed. (which is what you're getting at)

I believe that one of the members here caught an issue with some playtech games around the beginning of 2012 as the reel order of the free version of one of their games was not matching up with the reel order of the real play version.

Tying all this back to the original point - I don't believe that the major providers are changing the RTP of games when players change their coin size as at that point they would be playing a fundamentally different game, however if they were going to do this it's far more likely they'd offer a higher payout at higher coin sizes than a lower (higher RTPs encourage players to play for longer which is what they want high rollers to do!!).

 

[brianmon]

My point was that video slots exist only as programs and yes, while originally the math was based on a reels set with defined possibilities people often make the mistake of 'seeing is believing' and assume the reels they see are always tied to the event as opposed to video graphics generated after a press has ordered a RNG result. So if I implemented one of your changes above to the result probabilities the RNG would pick from the amended pool but the reels/graphics would be exactly the same to the player, no need to change them! All you need to do is make a tweak to the math, not the reels. Of course if the player could get hold of the reel maps, run a couple of million spins through then maybe they would notice they are getting an unrealistic lack of 3x9's but we all know that won't happen!

Agreed.
IGT use (or at least used to use) par sheets (virtual reels) to generate the wins.
And their RPT range is because they have different RTP settings.

I do remember someone from IGT commenting in a thread about 3 years ago. I don't have time to search for the thread, so I'll have to paraphrase...
'That depends on which maths model the customer orders'

I did have an article specifically about IGT par sheets. but i'm unable to locate it at the moment. However, there is a section on virtual reels here
xxx.slotdesigner.com/wp/wp-content/uploads/Elements-of-Slot-Design-2nd-Edition.pdf

If anyone's interested in reading up on the subject

 

[BlaCGaming]

Very interesting, even though I am having trouble to keep a clear head

One thing I wonder, how come some slots can be "capped" in wins? I have heard people talking about slot features (I think it was South Park) and that they could have doubled up again but the win was capped at 5000x Bet (or €, not sure).

Aren´t they putting a block to possibly higher wins? How is that justified?

Another thing I wonder about features, when you win for example 10 free spins x3 and you start them, is the whole free spins session already pre-determined? Or only each spin (talking about the scenario when you get disconnected after a spin has started).

Btw, is there a way to check all the possible outcomes of a slot accompanied with the chance of hitting it? For example a fullscreen wilds on DoA being the rarest hit?

Thanks again for all this great insight!

 

[dunover]

Very interesting, even though I am having trouble to keep a clear head

One thing I wonder, how come some slots can be "capped" in wins? I have heard people talking about slot features (I think it was South Park) and that they could have doubled up again but the win was capped at 5000x Bet (or €, not sure).

Aren´t they putting a block to possibly higher wins? How is that justified?

Another thing I wonder about features, when you win for example 10 free spins x3 and you start them, is the whole free spins session already pre-determined? Or only each spin (talking about the scenario when you get disconnected after a spin has started).

Btw, is there a way to check all the possible outcomes of a slot accompanied with the chance of hitting it? For example a fullscreen wilds on DoA being the rarest hit?

Thanks again for all this great insight!

You are referring to 'gamble' feature? That will always be capped for obvious reasons as it would skew the RTP way out of its range if not. Then again, I have messed around with gamble features in free mode like losing deliberately big feature returns on say Cashapillar by MG, and then seen nearly every time the game play 'very nicely' afterwards, like it was compensated in some way (of course that CAN'T happen on random slots ahem!) but that's another story.

The free games LOOK random when they spin, i.e. assuming no reel enhancements the spins appear exactly like in the base game i.e. Raging Rhino but they are pre-determined pretty certainly. If you play Thunderstruck 2 for example, and take the 5x spins for the first 8 bonuses you land, you WILL get a 5-of-a-kind win involving a wild on one of them paying 30-200x stake, something you wouldn't definitely get in the base game in 80 spins (wins over 30x are like rocking horse shit in the base game!) therefore it suggests the bonus trigger is simply generating a video representing a predetermined total say 194x. So the bonus (like the shitstorm feature) is simply a graphic that is the gateway to overall wins x stake not available in the basic pay tables, but which form about 20% of the game's overall RTP.

DoA is 100% predetermined, you can tell by the gimped way the wilds land, see the winner videos. People in one spin get 3 Wilds on the same payline, then a fourth conveniently lands and so-on. There have been bonuses with 11-12 wilds before. I worked out with a 5-scatter trigger on DoA the game is theoretically capable of paying over 50k x stake but there have been only one or two documented at over 10k. The value for a 50k x stake win isn't even in the pool of values, I doubt even 20k x stake is.

RTG also have feature limits - say on Rain Dance you get a 5 scatter trigger for 100 FS and most times it will retrigger and get you to the limit of 50,000 coins win 2500x stake or whatever it is - so the 5 scatters are simply a substitute for a </=2500x stake win that's not in the pay tables. The resulting hundreds of free spins are just a load of eye-candy BS to display the win.

 

[brianmon]

DoA is 100% predetermined, you can tell by the gimped way the wilds land, see the winner videos. People in one spin get 3 Wilds on the same payline, then a fourth conveniently lands and so-on. There have been bonuses with 11-12 wilds before. I worked out with a 5-scatter trigger on DoA the game is theoretically capable of paying over 50k x stake but there have been only one or two documented at over 10k. The value for a 50k x stake win isn't even in the pool of values, I doubt even 20k x stake is.

My biggest (x stake win) was only just over 10,000x, and that was 3 wildlines from 13 wilds. It 'conveniently' managed to miss 6 other possible wildlines.

Even if that had been triggered from 5 scatters. The win would only have been an extra 2,500x more - 12,684x stake

 

[PatrickNess]

(...) is there a way to check all the possible outcomes of a slot accompanied with the chance of hitting it? For example a fullscreen wilds on DoA being the rarest hit?

If you don't have the full source code of the game (RNG, reel mapping etc) or an offline version of the game (uncapped in spins/second) to do your own 100 billion simulations, the answer is no.

 

[slotter999]

One thing I wonder, how come some slots can be "capped" in wins? I have heard people talking about slot features (I think it was South Park) and that they could have doubled up again but the win was capped at 5000x Bet (or €, not sure).

Aren´t they putting a block to possibly higher wins? How is that justified?

Thanks again for all this great insight!

Capped wins are required to ensure a game stays within the boundaries of it's defined RTP. If you take the southpark example and the bonus game where you keep picking stuff behind the bushes then because there is no defined end to that activity you could theoretically sit there successfully clicking away into the millions and that would have to be compensated for in the game PAR sheet. Or...you just set a max win and voila the problem goes away. One of the reasons this is also required is due to the fact that this particular bonus is random at the point of choice so max win is a bit of a trade off for having random choice.

Interestingly, jack and the beanstalk has no such max win cap and in theory could also keep going for a week giving you the harps but my guess is the wilds are so heavily weighted not to land the chances of that actually happening must be silly because as far as I know there hasn't been much beyond x 10000 on that slot. It would be interesting to know what figure netent have based their calculations on though.

As for the odds of a full screen of wilds on DOA then as someone else has already said without the PAR sheet you couldn't say with any certainty. Somewhere between very long and very very very long is my guess.