Probabilities

[cjb]

reels

I started giving Loaded a try to figure out the reels. This one may take more time. Seems like the only way I can put reel 1 together is by using combinations more than once. I spent hours watching the reels for more combinations but had to quit cause I was starting to see it cross-eyed. I'm using the same technique as TS. I'll try again later or tomorrow..sad that I have all this time to do this, think I need to get a hobby..besides slots

 

[Zoozie]

I'll try again later or tomorrow..sad that I have all this time to do this, think I need to get a hobby..besides slots

At least it is cheaper cracking them in fun mode than by playing them for real

 

[vinylweatherman]

Weighted.

Zoozie, I note that you do not consider it weighted on grounds of how the landing site for each spin is determined. What I am getting at is the fact that reels 4 and 5 have nearly twice the number of symbols per reel than the first two. This effectively weights the slot against hitting the big 4 and 5 of a kind wins as the probability of hitting Thor on any given winline on a particular reel decreases as the number of symbols increases (unless the increase is proportionate, Ie, twice as many Thors on a reel twice as long). The good thing about doing it this way is that it leaves it fully open to analysis. If the weighting was done with fixed length reels as for the typical 3 reel "classic" slot then no simulator could crack the payout without knowing the weights.

I also noticed something else when studying my movie, and I hope I am wrong (no, sod that, I hope I am RIGHT, because it would make the slots slightly predictable in terms of overall mood as in "hot" or "cold").

Firstly, I need to demonstrate that the movie loop of the spins really does show the actual reels in motion, with all symbols in place and the differential reel length fully demonstated, and then??

Well, the RAMifications are potentially explosive, and could STRIKE at the very heart of our current understanding of MG RNG interpretation by the software in generating the results. This leaves me on the HORNS of a dilemma, and I intend to HAMMER away at this so that the next time I get totally SHAFTed by Thunderstruck I will at least have THORt more about it and crossed the MOAT of ignorance

 

[Zoozie]

Zoozie, I note that you do not consider it weighted on grounds of how the landing site for each spin is determined. What I am getting at is the fact that reels 4 and 5 have nearly twice the number of symbols per reel than the first two. This effectively weights the slot against hitting the big 4 and 5 of a kind wins as the probability of hitting Thor on any given winline on a particular reel decreases as the number of symbols increases (unless the increase is proportionate, Ie, twice as many Thors on a reel twice as long). The good thing about doing it this way is that it leaves it fully open to analysis. If the weighting was done with fixed length reels as for the typical 3 reel "classic" slot then no simulator could crack the payout without knowing the weights.

We only disagree on words, not on facts. When I hear weigthed I think
of the 3 reel slots where the reels are very small and the spins are
not random given how the reel looks.

There is a big difference between these two for me. The first I consider a 'fair' game. You know all game mechanics, like when using a deck of cards in VP. In the second case (what I call weigthed) it is easy to manipulate payout also and I find games 'I know how works' more interesting.

Zoozie

 

[cjb]

Loaded

Sorry, Zoozie, I tried to put the combinations together for Loaded the same way I did Thunderstruck but it just isn't working. Maybe because it has more lines? I have the combinations if anyone else wants to try. Watched that damn slot for at least 10 hours to get them

 

[Zoozie]

Sorry, Zoozie, I tried to put the combinations together for Loaded the same way I did Thunderstruck but it just isn't working. Maybe because it has more lines? I have the combinations if anyone else wants to try. Watched that damn slot for at least 10 hours to get them

Maybe you can post the result(pieces) of your work here? I might give it a try by using screenshots while spinning, this way you can sometimes see 4 symbols that seems to be reliable (I checked thunderstruck against your reels).

 

[cjb]

here's what I used for symbols..W=wild,S=scatter,D=lady with long dark hair,P=pink hat lady,B=blond lady,O=guy with orange,G=guy with green hat,H=house,Y=yacht,C=car,M=music sign,R=record player.
reel 1=BYO,BMO,BMH,CHR,CBY,CGY,CPH,DMO,GMP,GCB,GYR,HSR,HYW,HRG,MHY,MOC,MPY,OCH,OCP,OCG,PYH,PHS,RDM,RGM,RBM,RGC,SRB,SRD,YWR,YRB,YHS,YOC,WRG.
reel2=BRO,BMG,BCG,CBR,CHD,CPY,CGY,DCP,DMO,DSR,GCH,GYH,GYD,GCB,HWY,HDS,HMB,MOR,MBC,MGC,MGY,ORH,ORG,OMG,RGC,RBM,ROM,RHM,PMO,PYO,SRB,YOR,YDC,YPM,YHW,WYP.
reel3=BMH,BYR,COY,CGY,CWY,COR,DCW,DCO,HCG,HMB,HCD,HCO,HPR,GHP,GYD,GMO,MOS,MHC,MBY,MOS,OYG,OSR,ORH,PMH,PRB,RGM,RBM,RHM,SRP,YRG,YDC,YGH,WYD.
reel4=BYM,BCO,BGR,COR,CHY,CDM,CWY,DHC,DMO,GCD,GYM,GYD,HYB,HPM,HMG,HCW,MGY,MOR,MPR,MBR,ORH,ORB,PRS,PMB,RBY,RSC,RGC,RHM,SCH,YDH,YBC,YHP,YMP,WYH.
I didn't finish reel 5 but will if you get anywhere with the rest of these. By the way, what do you all mean when you say a slot is "weighted"?

 

[soflat]

Very interesting thread and results.

So which games would you expect to have payouts over 96% (since that is the average reported by the auditors)? Wouldn't it be the most popular ones, which are also the newest?

 

[Zoozie]

here's what I used for symbols..W=wild,S=scatter,D=lady with long dark hair,P=pink hat lady,B=blond lady,O=guy with orange,G=guy with green hat,H=house,Y=yacht,C=car,M=music sign,R=record player.
reel 1=BYO,BMO,BMH,CHR,CBY,CGY,CPH,DMO,GMP,GCB,GYR,HSR,HYW,HRG,MHY,MOC,MPY,OCH,OCP,OCG,PYH,PHS,RDM,RGM,RBM,RGC,SRB,SRD,YWR,YRB,YHS,YOC,WRG.
reel2=BRO,BMG,BCG,CBR,CHD,CPY,CGY,DCP,DMO,DSR,GCH,GYH,GYD,GCB,HWY,HDS,HMB,MOR,MBC,MGC,MGY,ORH,ORG,OMG,RGC,RBM,ROM,RHM,PMO,PYO,SRB,YOR,YDC,YPM,YHW,WYP.
reel3=BMH,BYR,COY,CGY,CWY,COR,DCW,DCO,HCG,HMB,HCD,HCO,HPR,GHP,GYD,GMO,MOS,MHC,MBY,MOS,OYG,OSR,ORH,PMH,PRB,RGM,RBM,RHM,SRP,YRG,YDC,YGH,WYD.
reel4=BYM,BCO,BGR,COR,CHY,CDM,CWY,DHC,DMO,GCD,GYM,GYD,HYB,HPM,HMG,HCW,MGY,MOR,MPR,MBR,ORH,ORB,PRS,PMB,RBY,RSC,RGC,RHM,SCH,YDH,YBC,YHP,YMP,WYH.
I didn't finish reel 5 but will if you get anywhere with the rest of these. By the way, what do you all mean when you say a slot is "weighted"?

I tried to crack the reels from your info. But I ran into problems trying to crack reel 1 and reel 2 and started checking your data.

On reel 1 I have problems with:BMO(I found it too) and BMH(I found it too), this should mean there would be 2 combinations with xBM and yBM, but you have only found RBM (I found it too). I am still looking for the last one (could it be that RBM occours twice on the reel?)

BMO


On reel 2 you miss: GBD, BDC (just want I found so far)
And you write DMO is on reel 2, I cant find yet, actually the sequence DM cannot be found. I think you mean BMO instead



If I can get the correct 3-symbols combinations on a reel, I think I can crack it.

Zoozie

 

[Zoozie]

YAY!

Reel 1 of 'Loaded' cracked.

The problem is 3 of the 3-symbol combinations occour twice.
I have used () around those that I have 'made up'. But when solving the
puzzle it is obvious that this is the only solution, so this is how the reel must be.

Chaining the ends give 3 fragments:
OCG-CGY-GYR-YRB-(RBM)-BMH-MHY-HYW-YWR-WRG-RGM-GMP-MPY-PYH-YHS-HSR-SRB-RBM-BMO-MOC
OCH-CHR-HRG-RGC-GCB-CBY-BYO-YOC
OCP-CPH-PHS-(HSR)-SRD-RDM-DMO-(MOC)

So how do you fit these fragments together as their are 2 solutions. Luckly
it is not matter at all! Max win/payback% will still be excacly the same(scream now if you disagree, I can prove this if needed).
So the reel is(have to put it on 3 lines):
OCG-CGY-GYR-YRB-(RBM)-BMH-MHY-HYW-YWR-WRG-RGM-GMP-MPY-PYH-YHS-HSR-SRB-
RBM-BMO-MOC-OCH-CHR-HRG-RGC-GCB-CBY-BYO-YOC-OCP-CPH-PHS-(HSR)-SRD-
RDM-DMO-(MOC)-(back to start)

which short is: CGYRBMHYWRGMPYHSRBMOCHRGCBYOCPHSRDMO

So reel 1 has 35 symbols. As I posted above there is an error with data for
reel 2 atm. I will try do reel3,4,5 later.

Now I understand how time consuming this is, as CJB complained about

But it would be a good idea if someone else try to solve the puzzle for
reel 1 as I believe the data posted by CJB is correct. I almost sure they
get excacly the same result I do, and will be a good test. (and is a fun puzzle)

Zoozie

 

[cjb]

cracked

Wow, Zoozie, that's great!!! I thought you would need to use some of the combinations twice. I tried that and wasn't sure if it would be correct but if you say it will work, I believe you. TS was a snap compared to this one.

 

[Zoozie]

I just found RPM on reel 3.And I do not believe HCD is there (probably you misread it instead of HCO from your notes)

Using this the solutions seems to be. (So I think reel 3 is cracked now).

DCW-CWY-WYD-YDC-DCO-COR-ORH-RHM-HMB-MBY-BYR-YRG-RGM-GMO-
MOS-OSR-SRP-*RPM-PMH-(MHC)-HCO-COY-OYG-YGH-GHP-
HPR-PRB-RBM-BMH-MHC-HCG-CGY-GYD-(YDC)

Short:
CWYDCORHMBYRGMOSRRPMMHCOYGHPRBMHCGYYD

Again I had to introduce two 3-symbol combinations to occour twice.
(statistical methods which will take a damn long time can show if this actually is true)

 

[Zoozie]

Reel4:

I have a problem with RGC, dont think it exists or missing a xRG combination.

 

[cjb]

reels

Now you know why I quit doing it..it was driving me crazy. I think RGC is correct. There is a GCD and I miswrote BGR..it should be BRG. I don't think I'll ever play this slot after all this...

 

[Zoozie]

BRG found on reel4.. back to cracking...

Edit: problem with BGR. no GRx found. almost sure BGR does not exist
also sure GYM does not exists. (but GYD does)

This gives reel4 as:

ORH-RHM-HMG-MGY-GYD-YDH-DHC-HCW-CWY-WYH-YHP-HPM-PMB-
MBR-BRG-RGC-GCD-CDM-DMO-MOR-ORB-RBY-BYM-YMP-MPR-PRS-
RSC-SCH-CHY-HYB-YBC-BCO-COR

So reel 4 cracked... 2 more to go

 

[Zoozie]

Reel 2 cracked:

GCB-CBR-BRO-ROM-OMG-MGC-GCH-CHD-HDS-DSR-SRB-RBM-BMG-MGY
-GYD-YDC-DCP-CPY-PYO-YOR-ORH-RHM-HMB-MBC-BCG-CGY-GYH-YHW
-HWY-WYP-YPM-PMO-MOR-ORG-RGC

giving the sequence:
CBROMGCHDSRBMGYDCPYORHMBCGYHWYPMORG

Reel 5 still to go.

 

[Zoozie]

Reel 5 cracked ( was the hardest).

MPC-PCD-CDY-DYO-YOR-ORH-RHM-HMB-MBC-BCH-CHY-HYW-YWR-WRH-RHM-HMD-MDC-DCO-COH-OHY-HYB-YBS-BSR-SRG-RGM-GMP-MPC-PCG-CGY-GYR-YRH-RHM-HMP

giving the sequence:
PCDYORHMBCHYWRHMDCOHYBSRGMPCGYRHM

I will put the reels for 'Loaded' into the simulator. It will be very interesting what comes out of it.

 

[Zoozie]

Finally! Here are the results you have been waiting for. Remember that these numbers
are only correct, if we are using the correct reels. However even a small mistake on a
reel can sometimes change payout% a lot, but since these data here seems realistic,
it is an indication that we got it right.

Again I have included the full program. Unzip the archive into a folder.
Then start 'start_simulation_LOADED.bat' or 'start_simulation_THUNDERSTRUCK.bat'

Here are some of the data, just ask if you want to know something else as well.
To get the payout% more accurate I need to run the program for some hours, or you can do
it yourself . Go ahead and do a few billions autospins.


Payout%~93.6% +- 0.2%
I had hoped for more here since this is one of the newer slots and could
give a clue about the payout% for other new/future MG slots.


Payout% without free spins~65.85%
Payout% if retrigges allowed ~101.4% (guess this is why they dont)
Average feature win (in bet size)~31.6 (very close to Thunderstuck)

Highest win combination:
H H W C H
Y W Y W Y
W Y D Y W

Total Win:10620
scatter win:0
Line 1:180
Line 2:24
Line 3:0
Line 4:0
Line 5:7000
Line 6:0
Line 7:0
Line 8:0
Line 9:24
Line 10:8
Line 11:0
Line 12:0
Line 13:180
Line 14:240
Line 15:2400
Line 16:180
Line 17:0
Line 18:0
Line 19:180
Line 20:0
Line 21:24
Line 22:180
Line 23:0
Line 24:0
Line 25:0


Zoozie

 

[Zoozie]

Ups. In the last archive Thunderstruck was set to no retriggers.
Here it is again. (I can not edit the old one, is too late)

Anyway you was probably playing with Loaded, and that was ok.
Running the simulation for 30 minutes gave payback% to 93.7%.


Zoozie

 

[GrandMaster]

I also got 93.7%. I also had the same problem as before. The Java programs runs OK on my old PC, bot on the new one or my laptop. Great job, Zoozie and cjb. I did not expect tha slots could provide so much intellectual stimulation.

 

[managra]

I also got 93.7%... I did not expect that slots could provide so much intellectual stimulation.

LOL, what a grand comment.

Am going to play devil's advocate a little. My apologies if after 15 pages of this thread i have lost track a bit of its original premise and if i'm entirely missing the boat.

Although these probability tests are very interesting, why do we bother? We KNOW that the payouts are not in players' favor but we continue to play slots anyway. Why? Because we hope to get REALLY lucky and beat the system. Same reason people keep on playing the lottery.

I know i'm ahead in Thunderstruck, Tomb Raider, and also Good To Go. The last game is the easiest to explain: the first day i played it, i was awfully lucky, hit a number of terrific free spins. I find it actually a very boring game with irritating game sounds. So, i seldom play it anymore. Might test it every now and then for 5-10 spins at low bets and since i have never been lucky again after that first day i quickly move on to other games. So, it would take an awfully long time to move down to break-even on this game.

As to Thunderstruck and Tomb Raider. I start off with lowish bets. If nothing happens after 10 spins max, i move on and try again later. I stay with these games if i hit 2 scatters a lot. It often results in getting the 3 scatters and the free spins. Once i make a nice hit i increase the bets, look at my new credits and set a new lower limit below which tells that the good run has come to an end and it's time to leave that game. So, i try to maximize the wins and minimize the losses.

It's probably true that on the long run those payout percentages will prevail but until then i'll count on luck and discipline, and not getting addicted to the big bets you could easily afford on a super lucky day.

 

[Zoozie]

'Tally Ho' analyzed. Reels was done by TeddyFSB. Tally Ho uses
excacly same reels as Thunderstruck. It just have another paytable, #freespins and freespin-multiplier. Here is the result:

Average feature win(in bet size):31.8
Payout%=96.0%+-0.1%(95.9%-96.1%) Can be found more accurate if needed

So the payout% is the highest found so far. Thunderstruck was only 95.0%.
TeddyFSB had a feeling Tally Ho would be better, and he was right.


Anyone know if Beach-party/Ladies Night also are 'reel-isomorphic' to Thunderstruck? Then I can analyze these two fast also.

Tally Ho simulation added. Extract and click the wanted simulation.

Zoozie

 

[Simmo!]

Hi Zoozie...i think what you're doing is great - its fun and interesting

Just a thought though: do the "free spins" % return probabilities you are doing take into account that each "free spin" doesn't require a wager - ie...its free?

Cheers

Simmo!

 

[Zoozie]

Just a thought though: do the "free spins" % return probabilities you are doing take into account that each "free spin" doesn't require a wager - ie...its free?

The program is Open Source , but here is the code-snippet:


if(freeSpinsLeft>0){
freeSpinsLeft--;
multiplier=this.getFreeSpinMultiplier();
}
else{
totalPay+=this.getNumberOfLines();
}

If there are freespins left, I do not add the spin to totalPay.

In the start when I made it, I made the simulation log every spin with
current line wins/scatter win/bankroll/freespins left/multiplier to check this and other stuff. I do consider it 100% accurate if the reels are correct.

Zoozie

 

[Slotmachine]

Wow! This thread has make giant leaps while I've been away. Great job Zoozie

Cheers,
SM