Cracking the slot reels - Here you go!

[gerilege]

It's also interesting that in case of slots like that, it is possible to calculate the payout exactly, not just via simulations, because the number of possible combinations are not so high (248832 in case of 5 reel drive).

I made some quick calculations, and the results really amazed me if I'm correct and the above reels are correct:
Whatever the return percentage of this slot is, 78.125%point comes from scatter wins. Am I correct in the below calculations?
P(Exactly 3 scatters)=BINOMDIST(3;5;1/4;FALSE)=0.087890625
P(Exactly 4 scatters)=BINOMDIST(4;5;1/4;FALSE)=0.014648438
P(Exactly 5 scatters)=BINOMDIST(5;5;1/4;FALSE)=0.000976563

Win (3 scatters) = 0.087890625 * 5 = 0.439453125
Win (4 scatters) = 0.014648438 * 20 = 0.29296875
Win (5 scatters) = 0.000976563 * 50 = 0.048828125
Sum = 0.78125

So if we assume that the overall payback is 95%, than 82.2% of the total payback would come from scatter wins...

I worked on this issue, and although I have not completely checked my program written for counting all possible combinations (not just for scattersm but also for line wins), I'm quite condifent that either the posted reel sequence is incorrect or the reels are weighted in case of 5 reel drive. Otherwise the payout would be too high.

 

[Zoozie]

Small update for you folks.

The reels we once cracked for Thunderstruck and Tally Ho 1 year ago were correct. The reels Kimss provided are identical. (Actually there 2 reel fragments was switched on reel 5, but this does not matter for payout and both choices was possible from the reel puzzles).

Also we now know that Thunderstruck is the same that it was 1 year ago!

I will work a little more on it and upload it later. HO HO HO is will be included and is a little interesting because it is a very low variance slot.

Zoozie

 

[Zoozie]

It's also interesting that in case of slots like that, it is possible to calculate the payout exactly, not just via simulations, because the number of possible combinations are not so high (248832 in case of 5 reel drive).

I made some quick calculations, and the results really amazed me if I'm correct and the above reels are correct:
Whatever the return percentage of this slot is, 78.125%point comes from scatter wins. Am I correct in the below calculations?
P(Exactly 3 scatters)=BINOMDIST(3;5;1/4;FALSE)=0.087890625
P(Exactly 4 scatters)=BINOMDIST(4;5;1/4;FALSE)=0.014648438
P(Exactly 5 scatters)=BINOMDIST(5;5;1/4;FALSE)=0.000976563

Win (3 scatters) = 0.087890625 * 5 = 0.439453125
Win (4 scatters) = 0.014648438 * 20 = 0.29296875
Win (5 scatters) = 0.000976563 * 50 = 0.048828125
Sum = 0.78125

So if we assume that the overall payback is 95%, than 82.2% of the total payback would come from scatter wins...

Right on! Payout is 248% so I it is obvious this slot can not have uniform reels probabilites. Ie. reels are weighted OR wins are determined by other means and the reels are just showing positions matching the win amount.

This is a little shocking news for meand I am disappointed MG did this short-cut which is against the slot-ethics for video slots.
I have never played these simple 5 reels slots, but this weighting should be obvious. Ie. the middle reel has 1 wild and total 12 positions. So it should be visible 3/12=1/4 of the spins on each of these reels, and I do not think it is so. Also Kimss reported to me that he had a hard time hitting some positions on the reels, and maybe it was on these slots.



Analyzing 5 Reel Drive , iterating through all combinations to calculate payout%.
************** Analysis complete *************
Seconds running:1.1
Number of iterations:248832
Total cost:248832
Total added wins:641848
Number of feature starts:0
Base payback(no free spins):2.579443158436214
Probability for getting a feature each spin:0.0
Average number of spins each feature with retriggers (calculated):0.0
Average number of win each feature(in bet size):0.0
Total payout (excact!)= 2.579443158436214~ 257.94%

 

[gerilege]

Right on! Payout is 248% so I it is obvious this slot can not have uniform reels probabilites. Ie. reels are weighted OR wins are determined by other means and the reels are just showing positions matching the win amount.

This is a little shocking news for meand I am disappointed MG did this short-cut which is against the slot-ethics for video slots.
I have never played these simple 5 reels slots, but this weighting should be obvious. Ie. the middle reel has 1 wild and total 12 positions. So it should be visible 3/12=1/4 of the spins on each of these reels, and I do not think it is so. Also Kimss reported to me that he had a hard time hitting some positions on the reels, and maybe it was on these slots.



Analyzing 5 Reel Drive , iterating through all combinations to calculate payout%.
************** Analysis complete *************
Seconds running:1.1
Number of iterations:248832
Total cost:248832
Total added wins:641848
Number of feature starts:0
Base payback(no free spins):2.579443158436214
Probability for getting a feature each spin:0.0
Average number of spins each feature with retriggers (calculated):0.0
Average number of win each feature(in bet size):0.0
Total payout (excact!)= 2.579443158436214~ 257.94%

Yes, I was almost sure yesterday when I checked the scatter wins that something was odd. Today I did not include the wilds when calculating the line wins (I also had to do the normal work for which I get paid ), and it was already over 100% point, plus the 78% from the scatter wins, already 178% at least, crazy. Tomorow I will add the wilds, so we can double-check our results, but 258% seems quite reasonable for me.

Needless to say, I'm also very disappointed, because MG was consistently stating that their video slots did not use such tactics.

 

[winbig]

It would be very interesting to see if they are weighted the same for free play.

 

[Slotster!]

It would be very interesting to see if they are weighted the same for free play.

This WOULD be interesting!

I'm personally entirely convinced that MG are identical in free as they are in real mode, but a bit of analysis would be good!

P.S - *SUPERB* thread, a real gem. Thank you!

 

[winbig]

This WOULD be interesting!

I'm personally entirely convinced that MG are identical in free as they are in real mode, but a bit of analysis would be good!

P.S - *SUPERB* thread, a real gem. Thank you!

As many bonus rounds/free spins that pop up so freely in free play, I think they do throw a few extra scatters in when you're playing for free. Not many, but just enough to make a difference.

An interesting argument to this would be that these extra scatters would block more wins, so it still seems as if the slot plays just like it does for real money.

Just like Hitman last night. I hit 4 bonus rounds in about a half an hour....Can you say the same for real play? I doubt it

I agree, great thread!

 

[Slotster!]

Just like Hitman last night. I hit 4 bonus rounds in about a half an hour....Can you say the same for real play? I doubt it

Sometimes, yes! There's even a bug when you hit it twice in the free spins... I'd love to see some work on free versus real - I'm totally convinced MG is fair. Now as for the rest of them...

 

[brianzz]

Just FYI, the B/M casino game makers run the programs for 10,000,000 games to attain the correct average long term return of a certain set of virtual reel strips. There is a bit of varience to the claimed returns in the short term, see example below.

Simplified example of a 3 reel game, Double Diamond

90% CONFIDENCE VALUES VOLATILITY INDEX = 10.476
HANDLE PULLS LOWER% UPPER%

1000. 59.46 125.71
10000. 82.11 103.06
100000. 89.27 95.90
1000000. 91.54 93.63
10000000. 92.25 92.92

That would probably help explain some of the outrageous returns you get when running a simulation.

 

[Zoozie]

Just FYI, the B/M casino game makers run the programs for 10,000,000 games to attain the correct average long term return of a certain set of virtual reel strips. There is a bit of variance to the claimed returns in the short term, see example below.

Simplified example of a 3 reel game, Double Diamond

90% CONFIDENCE VALUES VOLATILITY INDEX = 10.476
HANDLE PULLS LOWER% UPPER%

1000. 59.46 125.71
10000. 82.11 103.06
100000. 89.27 95.90
1000000. 91.54 93.63
10000000. 92.25 92.92

That would probably help explain some of the outrageous returns you get when running a simulation.

Yes, but I did not run an simulation. I analyzed it finding the exact payout (which can be expressed as a fraction).

The analyzer can also simulate and here is the result after 1M and 10M:

Iterations:1000000
Payout%:2.5461415555555558~ 254.61%
---------------------------------------------------
Iterations:10000000
Payout%:2.579288688888889~ 257.93%
---------------------------------------------------

 

[Zoozie]

It is also weird that 5-reel drive has two identical symbols but with different reel-identification numbers.

Kimms, you have any idea why and you are of course sure you did it correct?

 

[gerilege]

It is also weird that 5-reel drive has two identical symbols but with different reel-identification numbers.

I think they solved the "wilds only on reel 2-4" problem this way. For some reason they wanted the reels to be equal length, so they replaced the wilds with an identical symbol on reel 1 and 5.

 

[Zoozie]

I think they solved the "wilds only on reel 2-4" problem this way. For some reason they wanted the reels to be equal length, so they replaced the wilds with an identical symbol on reel 1 and 5.

In the newer 5-reel slots same symbol can be multiple times on each reel and they could just have done the same here. Ie. one symbol count be twice on reel 1 and reel 5. But since there are only 1 instance of each symbol (by ID) on each reel your explanation is very good.

MG probably had no expertise (or even just slightly skilled math. people) at that time when the first 5-reel slots was developed. And they probably used weighted reels which they already had done in the 3-reel slots and they just generalized it. Later they realized they had screwed up by misunderstanding the whole idea behind a 5-reel slot. And soon they became the leading casino software provider and got it right. Designing/balancing the slots is a little hard and MG really did a good job. (Thunderstruck clones etc.). When Rival software saw the light a little over 1 year ago, they had massive problems balancing the slots and changed it regulary. (And I still believe Rival should increase feature payout and lower base-payout to compensate it)


I still believe that Boss software still 'cheats' this way with
their 5-reel slots and I am also not sure RTG does it correct. I am pretty sure Crypto got it right, and it would be interesting if someone can crack the reels on one of their standard slots. (freespin slots, no wierd mechanics - maybe the new Thor slot?)

Zoozie

 

[gerilege]

From another perspective

I forgot to post yesterday that I also have a straight mathematical proof that those 5 reel slots should be weighted with the given reelset, without any simulation or counting all possible combinations. It is enough to show that the payout would be significantly over 100% without weighted reels.

I have previously shown that the scatter payout would be 78.125%.

Let's add the five top symbols payouts, which is 10000 (1111.11 x total bet) for five "Wheels".

The total number of possible reel combinations is 248832. How many combinations would give five "Wheels"? We have:
- 9 combinations without a wild.
- 9 combinations with 3 wilds on reel 2-4.
- 3 times 9 combinations with 2 wilds on reel 2-4. (i.e. wilds are on reel 23, 24, 34)
- 3 times 9 combinations with 1 wild on reel 2-4. (i.e. wild is on reel 2, 3, 4)
That is a total of 72 combinations.
Now the chance for five "Wheels" would be 72/248832. And the payout is 10000/9 x total bet. So the return percentage for five "Wheels" is:

(72/248832)*(10000/9)=(8*10000)/248832 ~= 0,3215 = 32,15%.

So ony the five wilds + scatter wins would sum up to 110,275%.

 

[Zoozie]

3 new slots in the slot-analyzer

I have added Loaded, HO HO HO and Five Reel Drive to the slot-analyzer.
Five Reel Drive does not really work though, because the reels are weighted. But then we at least know that, so I will keep it there.

Also for the first time I saw the "Loaded" payout% and it was good news.
It has highest payout we have found for a MG slot yet. Also this is a much newer slot that those we already had analyzed (Thunderstruck and HO HO HO) indicating
that MG has not lowered payout on the newer slots.




************** Analysis complete *************
Seconds running:158.0
Number of iterations:46652760
Total cost:46652760
Total added wins:31171484
Number of feature starts:428760
Base payback(no free spins):0.6681594829544919
Probability for getting a feature each spin:0.00919045304072042~1/109
Average number of spins each feature with retriggers (calculated):24.0
Average number of win each feature(in bet size):32.071655181815615
Total payout (excact!)= 0.962912523841146~ 96.29%


I added the link to the slot-analyzer in my signature.

 

[steinhaug]

I am pretty sure Crypto got it right, and it would be interesting if someone can crack the reels on one of their standard slots. (freespin slots, no wierd mechanics - maybe the new Thor slot?)

Any TCP packet experts in here? I believe all reels in all games in all casinoes can be cracked easily - since I believe most likely same approach is used when feeding the data to the standalone installation. Atleast MG works this way, I do not think the flash and standalone system differs that much when you get down to the bone of it. As mentioned before, sniffing the packets getting to your machine on a spin - theese packets contains reelposition for each spin, what you won and such. You cant manipulate anything with the data - so I don't look at this as breaking an entry, however one can use this pin point data to accurately place the symbols on the reels. If all were correct you need like 50 spins and you have cracked the reels, however, as mentioned before some positions on the reels going this method is darn hard to get for some (???weighted???) reason...

PM me if anyone has experience with TCP packets.

 

[steinhaug]

It is also weird that 5-reel drive has two identical symbols but with different reel-identification numbers.

Kimms, you have any idea why and you are of course sure you did it correct?

Yes i'm very sure, and I can't be wrong either since I'm using MGs own mapping of the symbols,

Seems to me they once made a slot, and tweaked it for a new slot wothout reprogramming the whole slot. They had a symbol to much, and there you go? As a programmer that would by my instant assumption. All programmers are lazy,

 

[Zoozie]

So Isis and Dolphin tale are not yet in the flash casino?

The 5-reel slots are all weighted as can be seen easy from the reel like
Five Reel Drive. The 9-lines are clones of Tomb Raider or Thunderstruck.
So I will not implement Thunderstruck,SunQuest,Spring Break, Ladies Night etc.. since they are the same, but I will mention it in the slot-analyzer.
And then there is a crazy amount of Tomb Raider clones, and I have already done Tomb Raider, but it is not working before I implement the "bonus not multiplied during freespins".

It would be VERY interesting to get the MG progressives also. Are the in the flash casino? Then we can also see how high the jackpot has to be before payout exceeds 100%

Then there is also the standard slots with the new reel layout like
Flying Circus and Witch Doctor. They are not in the flash yet also?

Also the WagerWorks Cleopatra would be interested if anyone want to crack it the old fashioned way. But I think you can see the reels quite easy on cleopatra.

Zoozie

 

[steinhaug]

...progressives...

Sweet, I never saw this option in the flash casino... I'll get creacking right away,

 

[steinhaug]

Doh! I have no real accounts left, and I am not thinking of getting any more either. This means it's kinda impossible for me to crack theese reals since I can't get any spins on them. I did two however, but I miss some symbols since I must map from the startup layout. Maby some of you can finnish the job.

Major Millions:
0 - Scatter
1 - Major Millions progressive
2 - Officer saying hey
3 - ???
4 - Airplane
5 - Tank
6 - Top Secret
7 - Medals
8 - Hat
9 - Ammunition box
10 - Googles

Reel 1: 1,9,5,10,7,3,10,8,4,10,2,8,10,6,0,10,7,5,0,10,9,4,10,6
Reel 2: 1,6,9,3,10,4,8,2,9,5,8,3,7,6,10,0,8,7,4,8,5
Reel 3: 1,5,6,4,5,6,4,5,6,2,5,6,3,5,6,9,0,10,7,5,6,8,7,9,0,10,8
Reel 4: 1,9,5,8,3,10,6,8,3,10,5,8,2,9,6,8,4,7,9,0,8,7,4,9,8
Reel 5: 1,10,5,8,3,9,4,7,3,5,10,0,7,4,6,8,2,7,9,0,7,6

Mega Moolah:
0 - Leon
1 - Elephant
2 - Moskus/bison
3 - Giraffe
4 - ???
5 - ..Deer, tilted head with horns
6 - A
7 - K
8 - Q
9 - J
10 - 10
11 - Feature, Moolan

Reel 1: 0,10,9,6,5,7,8,3,10,7,4,8,9,5,10,2,9,6,5,7,8,3,10,7,4,8,9,1,10,11,6,9
Reel 2: 0,7,10,9,5,8,7,4,6,8,3,9,10,11,2,6,10,9,5,8,7,4,6,8,3,9,10,1,5,8,7,9
Reel 3: 0,9,4,10,6,2,7,8,4,9,7,1,8,10,5,7,8,3,6,2,11,10,5,6,2,7,8,4,9,7,1,8,10,5,7,8,3,6,1,9,3
Reel 4: 0,8,9,10,11,8,1,7,5,6,10,3,9,8,4,7,6,2,9,3,7,5,8,9,10,11,8,1,7,5,6,10,3,9,8,4,7,2,6,9,2,7,4,10
Reel 5: 0,6,3,9,10,5,8,9,1,7,8,2,10,7,8,4,7,9,5,6,11,3,6,9,3,10,5,8,9,1,7,8,2,10,7,8,4,7,9,2,10,4
(Reels are served twice, but they are identical - the other slots doesnt do this?)
Reel 1: 0,10,9,6,5,7,8,3,10,7,4,8,9,5,10,2,9,6,5,7,8,3,10,7,4,8,9,1,10,11,6,9
Reel 2: 0,7,10,9,5,8,7,4,6,8,3,9,10,11,2,6,10,9,5,8,7,4,6,8,3,9,10,1,5,8,7,9
Reel 3: 0,9,4,10,6,2,7,8,4,9,7,1,8,10,5,7,8,3,6,2,11,10,5,6,2,7,8,4,9,7,1,8,10,5,7,8,3,6,1,9,3
Reel 4: 0,8,9,10,11,8,1,7,5,6,10,3,9,8,4,7,6,2,9,3,7,5,8,9,10,11,8,1,7,5,6,10,3,9,8,4,7,2,6,9,2,7,4,10
Reel 5: 0,6,3,9,10,5,8,9,1,7,8,2,10,7,8,4,7,9,5,6,11,3,6,9,3,10,5,8,9,1,7,8,2,10,7,8,4,7,9,2,10,4

Missing symbols would be:
Mega moolah, reel 1 - What is the symbol after the K? This would be symbol 4.

Missing Major millions:
What is the symbol after the medals on reel1? This is symbol 3

 

[steinhaug]

OK, I did the last progressives, but I need a little help from you people to finnish of the reels. It's very easy work for you, Read comments in red.

Fruit fiesta:
0 - Fruit Fiesta
1 - ???
2 - Bar x3
3 - Ananas
4 - Water melon
5 - Orange
6 - Lime
7 - Plum
8 - Apple
9 - Bell
10 - Scatter

Reel 1: 0,9,5,7,3,5,7,1,0,8,4,6,7,2,6,7,10,6,8
Reel 2: 0,6,3,2,9,6,1,5,7,2,6,3,5,7,0,9,6,4,10,7,4,2,8,3,7
Reel 3: 0,9,6,3,8,2,4,8,3,5,7,9,5,7,10,4,7,1,4,6,8,4,7
Reel 4: 0,9,5,2,8,3,5,6,7,4,10,5,8,1,9,5,6,9,8
Reel 5: 0,5,8,10,5,6,2,7,9,3,6,4,7,9,6,1,4,10,9,3

Symbol 1: Whats the symbol after the plum on reel 1.

King Kashalot
0 - King
1 - ???
2 - Yellow knight
3 - Lady with red hair
4 - Steak/meat
5 - Chicken
6 - Fish and schrimps
7 - Pudding
8 - Plate of fruits
9 - Dragon
10 - ???

Reel 1: 0,5,8,7,6,8,7,4,8,7,1,6,8,7,3,5,7,6,10,5,7,8,4,6,8,2,5,6,8
Reel 2: 0,8,6,3,8,9,6,7,5,4,6,9,7,8,6,1,8,7,10,8,7,6,5,7,2,8,9,5
Reel 3: 0,6,7,5,3,8,5,7,4,9,8,6,1,5,10,8,6,7,9,8,2,7,8,6,7,8
Reel 4: 0,5,6,4,2,6,8,3,7,10,6,2,8,6,7,4,8,1,9,3,7,6,2,8,4
Reel 5: 0,8,2,6,4,3,10,1,4,7,3,6,4,5,2

Missing symbols:
1: Symbol after "pudding" on reel 1.
10: Symbol after "fish and schrimps" on reel 1.

Cash Splash:
0 - Scatter
1 - Cash Splash
2 - ???
3 - Grey 7
4 - Bar x1
5 - Cherries
6 - A
7 - K
8 - Q
9 - J
10 - 10

Reel 1: 1,10,6,8,4,6,8,2,9,5,7,8,3,7,8,0,7,9
Reel 2: 1,7,4,3,10,7,2,6,8,3,7,4,6,8,1,10,7,5,0,8,5,3,9,4,8
Reel 3: 1,10,7,4,9,3,5,9,4,6,8,1,10,6,8,0,5,8,2,5,7,9,5,8
Reel 4: 1,10,6,3,9,4,6,7,1,8,5,0,6,9,2,10,6,7,10,9
Reel 5: 1,6,9,0,6,7,3,8,10,4,7,1,5,8,10,7,2,5,9,10,4

Missing symbol:
2: Symbol after Q on reel 1.

Lotsaloot:
0 - Lotsaloot (text)
1 - Bucket of gold
2 - 7
3 - Bar x1
4 - firklver
5 - A
6 - K
7 - ???
8 - J
9 - 10

Reel 1: 0,6,8,3,7,6,5,7,1,8,9,6,4,7,2,9,5,7
Reel 2: 0,9,8,3,4,5,1,9,6,2,5,7,3,8,0,9,7,3,5,6,2,4,5,7,6,2,7,8,6
Reel 3: 0,9,7,4,5,1,8,4,3,7,8,0,9,6,3,4,7,2,5,7,4,8,6
Reel 4: 0,4,9,6,2,5,8,7,0,5,6,3,9,8,1,9,8,5,7
Reel 5: 0,4,6,3,9,8,2,5,3,9,7,0,4,6,8,1,9,5,7,6

Missing symbol:
7 - alternative a: Reel3, in between 10 and LotsaLoot.
7 - alternative b: Reel3, in between A and K

- - - -

Friom what I see there are no more slots to be mapped from the flash casino now, unless they release new games.

 

[gerilege]

R
Analyzing 5 Reel Drive , iterating through all combinations to calculate payout%.
************** Analysis complete *************
Seconds running:1.1
Number of iterations:248832
Total cost:248832
Total added wins:641848
Number of feature starts:0
Base payback(no free spins):2.579443158436214
Probability for getting a feature each spin:0.0
Average number of spins each feature with retriggers (calculated):0.0
Average number of win each feature(in bet size):0.0
Total payout (excact!)= 2.579443158436214~ 257.94%

I have also written an analyzer for the 5 reel drive, and I got slightly different results (275% payout) than you. I ported my c++ program to JAVA for your convinience, and attached it to this post. Could you please have a look whether I made something wrong? It would be very nice if my program could come to the same payout%.

After that I plan to collect some spin data in free mode (1-2000 spins), so I would have some data available about the possible reel weighting. Then I would extrapolate this weighting (use it as input for the simulation), so we could have an approximate payout percentage simulated with the weighted reels.

What do you think?

 

[Zoozie]

I have also written an analyzer for the 5 reel drive, and I got slightly different results (275% payout) than you. I ported my c++ program to JAVA for your convinience, and attached it to this post. Could you please have a look whether I made something wrong? It would be very nice if my program could come to the same payout%.

What do you think?

Nice job. Still we have a little difference that should be found.

Maybe you can use the following log info from my program:
Number of iterations:248832
Total cost:248832
Total added wins:641848

You have to add all the wins also I think, and what did you get?

 

[Zoozie]

I found it.

One of us got the paytable for 4 of the top-pay-symbol wrong. I have it at 500 in my simulator while you have it at 1000.

And if I change mine to 1000 I get:
payout=275,63%, just like you did!

I am at work and I can not check the paytable.

 

[steinhaug]

After that I plan to collect some spin data in free mode (1-2000 spins), so I would have some data available about the possible reel weighting. Then I would extrapolate this weighting (use it as input for the simulation), so we could have an approximate payout percentage simulated with the weighted reels.

I am doing this now, but on the feature slots. I have a feeling theese are weighted aswell since I know certain reel positions just never happend. I will be collecting a few 1000 spins on several machines, and maby some of you people can program a little piece of software to calculate between the reel positions to see if its random or not? This would be a job for maby you and zoozie?

I'll get back to you later with the data from this, you will get the data like this for each spin:
0,5,10,4,5
10,12,0,0,4
20,18,4,2,1
...

Theese are the reelpositions for each spin, I would think that if reel 1 has 33 symbols, that means reelposition may stop from 0-31. In a 1000 spins we should expect pretty much all positions to spread equaly?

Anyways, You'll here from me a little later when I have accumulated and formatted the data for you. If you have any special games you would like data from (within the flash casino that is) please tell me and I'll start collecting some thousand spins for you.

Great work!