Slot Machine?

[LarryG]

How self-aware do you want to be with it? That is, is it kind of a humorous thing, or is it supposed to be very canon? I doubt that any casino owner would use religious symbols in a slot machine.

With that said, there are glyphs, the pagan symbol, the hammerite symbol, the mechanist symbol, and the keeper symbol. And a burrick-head.

I want to be sensitive to the canon and not humorous.

Since religous symbols should be out, I guess that rules out the hammer and the gear. Glyphs are only known to the Keepers, so they are out, as is the keyhole symbol of the Keepers. So that leaves, what as possibles? Sword, Mace, Shield, Wand, fruit, Money (or purse), animals, ... The trick is that the symbol needs to be recognizable limited to a 20x20 pixel square (or so), so it can't be complicated looking.

I'm open to suggestions (links to possible icons would be nice too!).

[LarryG]

The latest in slot-o-meter(tm) design, thanks to the CoSaS model whiich I cut apart, put back together, and retextured. I also removed the joint on the model which seems to have been intended to let you open up the machine for servicing (or looting, perhaps), but was never fully finished on the inside, so the interior is transparent in-game. In its place are the odometer rollers.

Now on to making a better looking handle-switch and rigging it for payouts. I'm thinking that Garrett should get as many plays as he can find special chips to play with. That should make this simpler than setting it up like a store where Garrett "buys" each play.

Any suggestions about the payout odds I should go for? And should the payout be in gold, or more of the special chips to play with? I'm leaning towardss the latter. Otherwise tha amount of loot in the game will vary each time the slots are played. What do you think?

[Digital Nightfall]

I also removed the joint on the model which seems to have been intended to let you open up the machine for servicing (or looting, perhaps), but was never fully finished on the inside, so the interior is transparent in-game.

You mean you never played and don't know?

It opens up when you successfully pick the lock on it and loot it. The inside was finished, though here finished means that no surfaces which could never be seen were included, to save on poly counts.

Nice work making it playable.

[scarykitties]

The following is a response to this in another thread.

Unfortunately the payout box is internal to the slot machine model. I'm treating it like a chute because the coins will fall through to the floor and be hidden otherwise. But it really was designed as a payout box if you look closely.

No, I meant just put a box below it on the ground for the coins to drop into.

5) figure out a way to get a payout on any triple and/or any double. Right now I took the easy way and pay if the rightmost two match only. Unfortunately I don't see an easy way to do pattern matching on qvars. Maybe NV or someone else will have a suggestion ... hint, hint.

I'm not sure exactly how the digits function, but I assume you've already considered some kind of CD for alike symbols and inverted CDs for nonmatching symbols, all going to various RATs that would handle the coin payout?

Maybe make each symbol on each dial send a number, and make a set-up like unlocking a door with a number code. That is, let's say that each dial has 5 different symbols. The first would send a "1" along whatever link. Second, 2, etc. Then, you would adjust payout for various number codes, such as 11xx, 22xx, etc.

The idea seemed to make sense a moment ago when I was lying in bed, but now that I'm up and typing it, it sounds a lot less solid. Oh, well.

[LarryG]

Still to do:

1) Require frobbing with a gaming chip to operate -- High priorityCOMPLETED
2) Time delay or RAT so that the slot-o-meter(tm) has time to complete its Tweqs before payout happens -- Med priority
3) Switch payout to gaming chips
4) impliment a grand payout for hitting all 4.-- High priority
5) figure out a way to get a payout on any triple and/or any double. Right now I took the easy way and pay if the rightmost two match only. Unfortunately I don't see an easy way to do pattern matching on qvars. Maybe NV or someone else will have a suggestion ... hint, hint.-- Med priority
6) I'd like to get the chips to deposit in the obvious place instead of having them drop to the ground, but I can't quite see how to do that either. Landing on the ground isn't all that bad, is it? -- Low Priority

Addressing the High Priority items first.

4) I believe I know how to do this for hitting all four. Trap>Quest Var: %1111:slot_combo should do the trick. The problem is with hitting any three. Fortunately #5 is medium priority, & I can ignore it for now. By the time I get to it NV may have enhaced the NVTrigQVar script to give me wild cards or something equivalent for digit matching of qvar values. I can hope.

1) I think that the way to do this is to turn the one armed bandit's one arm into a locked door, and to have the chips as keys. Right now I am using stdbutton on the lever to get it to auto-return after frobbing. But if I use stddoor, I should be able to lock it and get it to auto return. AND getting the chip to self destruct on use is part of stdkey. What I will need to do is make the chips stack in inventory and deliver up one at a time. I think the Use Ammo flag will let that happen ...

[qolelis]

5) figure out a way to get a payout on any triple and/or any double. Right now I took the easy way and pay if the rightmost two match only. Unfortunately I don't see an easy way to do pattern matching on qvars. Maybe NV or someone else will have a suggestion ... hint, hint.-- Med priority

Hmmm... Any even-digit number with all digits being the same is always dividable by 11. If you right- or left-shift the slot machine number and then test for 11-dividability, you could identify winning numbers... if my thinking is correct? No, it wasn't altogether correct, but it could be a start.

Sorry, I didn't read your post above.

6) I'd like to get the chips to deposit in the obvious place instead of having them drop to the ground, but I can't quite see how to do that either. Landing on the ground isn't all that bad, is it? -- Low Priority

Maybe you could PhysAttach the coin(s) in the right place. Just remember to remove the PhysAttach-link when the coin is frobbed. I've had situations where picking up a PhysAttach:ed object crashed the game.

Instead of PhysAttach you could set Location & Rotation for Physics->Model->Controls.

Another way could be to put a solid terrain brush just below the deposit area. You would of course need an extra solid for each slot machine. Using a solid, you can let the default physics do most of the work. Maybe...

[raph]

I'm completely not tech-savvy for stuff like that but can't the slot machine object from SS2 be imported into T2 with a different steampunk-looking model?
Apologies if it's been attempted already and I didn't see.

[Wille]

My suggestions for slot symbols:

-Gemstone from EP (ask NV for render)
-Lucky coin from EP (ask NV again)
-Apple from EP (NV)
-Owner's signature letters (i.e. R for Ramirez)
-Gemmed ring from EP (NV)

[LarryG]

... Instead of PhysAttach you could set Location & Rotation for Physics->Model->Controls ...

That was what I thought I would do. I was thinking that the gaming chips would have that set, and I would not let them be dropped, but kept in inventory. That way I don't have to worry about taking those controls off too. The higher denomination chips would have stack counts on them and convert to units when picked up ... That seems the easiest. We'll see.

My suggestions for slot symbols:

-Gemstone from EP (ask NV for render)
-Lucky coin from EP (ask NV again)
-Apple from EP (NV)
-Owner's signature letters (i.e. R for Ramirez)
-Gemmed ring from EP (NV)

Are you thinking of the object textures? Object textures may look nothing like the objects ... or is there some other "render" you have in mind ... ? I'm pretty much OK with my latest set if none of those pan out. Thanks, though.

[LarryG]

... I'm not sure exactly how the digits function, but I assume you've already considered some kind of CD for alike symbols and inverted CDs for nonmatching symbols, all going to various RATs that would handle the coin payout?

Maybe make each symbol on each dial send a number, and make a set-up like unlocking a door with a number code. That is, let's say that each dial has 5 different symbols. The first would send a "1" along whatever link. Second, 2, etc. Then, you would adjust payout for various number codes, such as 11xx, 22xx, etc.

The idea seemed to make sense a moment ago when I was lying in bed, but now that I'm up and typing it, it sounds a lot less solid. Oh, well.

The slot machine is a fancy lookin' odometer, is all. That means the thing is controlled by a single qvar (the one I use for this is named slot_combo). I'm going to have several working slot machines, maybe 10 or more, so the setup to check the qvar values can't be too complicated. Each test needs to have a unique QuestVarTrigger. The tests are in the form operator test_value :qvar_name The possible operators available are:

 
=
 Is true if the variable is exactly equal to the test_value.
 
<
 Is true if the variable is less than the test_value.

>
 Is true if the variable is greater than the test_value.

&
 Is true if any of the bits in the argument are also set in the test_value.(Bitwise 'and'.)
 
"
 Is true if the lowest decimal digits in the variable, the same number as there are in the argument, are the same as the test_value.

+ 
Is true if the variable has just increased by exactly the test_value. 

- 
Is true if the variable has just decreased by exactly the test_value.

} 
Is true if the variable has just increased by at least the test_value.

{ 
Is true if the variable has just decreased by at least the test_value.

% 
Is true if the variable can be evenly divided by the test_value.

Of these, % and & seem the best candiates. Unfortunately there is no left-shift or right-shift and compare operators, nor a masking comparator, nor multiple operators (left-shift followed by right-shift followed by >, as an example).

What would be nice wouild be if could use wild cards. For example, ? to represent any single digit and _ as any digit at all, then I could do something like =_???:slot_combo for one test, =?_??:slot_combo for a second test, =??_?:slot_combo for a thrid, and =???_:slot_combo for the fourth. That would allow me to trap on any triple showing up, just the way that %1111:slot_combo lets me test for any quadruple. Unfortunately, there are no wild cards enabled by the testing script. So I need to see if I can do anything simple with & ...

Another problem is the sheer number of payouts I need to make playing the slot machine fun. There are 10000 possible values that can show up. I want about 1 in 10 to pay a small amount, maybe 1 in 20, but that would mean at least 500 possible ways to win. I can't have that many discrete value tests. I need to use pattern matching of some kind. I want no more than 20 tests per slot machine ... even that is a LOT. Ideally I really want only 2: a grand payout test and a lesser payout test.

[scarykitties]

I'm guessing you've already considered each dial's value telling a blueroom AI to walk to one of *number of possible digits per dial here* different AIWatchObj markers so that the AI for each would stop and frob one of the possible buttons, with each button being linked to a series of possible RATs that would lead to payouts?

For instance:

Each blueroom would have four AI, each with:
button 1 (corresponding to symbol 1)
button 2 (corresponding to symbol 2)
button 3 (corresponding to symbol 3)
button 4 (corresponding to symbol 4)


button 1, 1, 1 and 0 -> RAT -> small payout
etc.
button 1, 1, 1, and 1 -> RAT -> large payout
etc.

The AI would have a time warp of, say, .1 and would be put on a patrol, which would be killed when the dial stopped turning, causing the AI to frob whichever button via AIWatchObj that it ended nearest.

I guess the problem there, as you pointed out, would be that each combination would need to be individually added.

It would be so easy if you didn't need those small payouts (only have one AIWatchObj-frobbed button on the patrol route).

[LarryG]

Unfortunately, I don't think that idea will work with the way that odometers are set to function. There is no easy way to extract the value of a position. That's what left- and right-shift operators would enable (combined with the & operator, I believe, or is it an OR operator I need ... I forget. I can look it up if I ever get a full set of operators ...) . But those operators are not available.

And small payouts are essential to making playing the solts fun/addictive.

[scarykitties]

Out of curiosity, does this feature prominently in the gameplay of the level, do you just have an idea you really want to try, or do you just have a thing for details?

Is the limitation of no wild cards/no multiple operations an odometer thing, or a script thing?

[LarryG]

It provides a "game-within-a-game" to play. Not an essential plot element in its own right, but it supports an essential plot element.

The operators are Qvar script specific. They are the ones allowed by NV's extention to the TrigQVar script, NVTrigQVar.

[qolelis]

That's what left- and right-shift operators would enable (combined with the & operator, I believe, or is it an OR operator I need ... I forget. I can look it up if I ever get a full set of operators ...) . But those operators are not available.

...but that's for extracting bits, right? Extracting digits can be done with the existing operators. Just use '%' and '/' (remainder and integer division).

[sterlino]

nice idea... i would like to do a version of mine of the thief's slotmachine.

should be simple: there is a way to let the odometer to create a random sequence. the problem that is still a lot of Qvartraps and questvartraps to place on the map.
If you want to know what i am talking about look at my old example: ''easy bank''
that mission is a demonstration of a random creation of codes for the odometer.

i will try to convert that mission in one for a totally functionally slot machine for thief !

but still wonder if i will be able to find another better and easiest way to make a odometer random sequences codes... ._.

[LarryG]

You're thinking of the TrapSetQVar operators. There is no remainder operator for TrigQVar.

But, and I'm thinking out loud here, NVTrapSetQVar lets me set another QVar based on the one that just got the random number, so I can isolate each digit into a separate QVar and then do RATs on combinations of them being equal ... I can test them being equal because NVTrigQVar lets me test QVars against each other for equality ... I think that may be the way to go ...

[Wille]

Are you thinking of the object textures? Object textures may look nothing like the objects ... or is there some other "render" you have in mind ... ? I'm pretty much OK with my latest set if none of those pan out. Thanks, though.

No I ment renders of the actual objects like NV showed us in this thread:

http://www.ttlg.com/forums/showthrea...roving&page=24

[LarryG]

Ah. Screen shots turned into icons. Maybe. I'll see if any of them work for this.

[sterlino]

just one thought under the shower... now i found a simple solution.
but i need to do some test.

tomorrow i will create my version of the 'slot machine'

[LarryG]

OK this is my thinking on setting up the QVars. First, each slot machine needs a separate set, otherwise they will all change to the same number at the same time. There are five QVars needed per. Here are the operations to load them.

1. set the base number randomly when the machine lever is pulled [d9999:slot1_combo]
2. store the base number into a new variable [+slot1_combo:slot1_dig1]
3. get the 1st digit via remainder [%10:slot1_dig1]
4. store the base number into a new variable [+slot1_combo:slot1_dig4]
5. shift the saved number right by dividing [/10:slot1_dig4]
6. store into a new variable [+slot1_sav1:slot1_dig2]
7. get the 2nd digit via remainder [%10:slot1_dig2]
8. shift the number right, again, by dividing [/10:slot1_dig4]
9. store into a new variable [+slot1_sdig4:slot1_dig3]
10. get the 3rd digit via remainder [%10:slot1_dig3]
11. get the 4th digit by dividing [/10:slot1_dig4]

That just gets me as far as having the independent digits. So that's 11 traps that have to be triggered in order for this to work. Then there is the work to do a pairwise comparisons.

If I just wanted to check for triples and and quadruples, life would be easier. The patterns are xxxx, xxxn, nxxx, xnxx, and xxnx. but that's only 41 hits out of 10000. Not enough. I think I need to add in hits for two pairs.

Sheesh! I sure could use a script to help with this ...

[sterlino]

i made my version of the slot machine... but watching the LarryG's version i think i will give up... i am now just curious about the model of Larry..

hey it's a single object or a combined solid + object ?
do you will release a demo version ?

[LarryG]

Two objects: the slot-o-meter(tm) slot machine and the slot machine lever which is PhysAttached to the slot-o-meter(tm) slot machine. The lever establishes the QVar values for the slot-o-meter(tm) slot machine, which is just an odometer after a bad acid trip.

I'm currently workjing on converting the levet to be a lockbox in a clever disguise with gaming chips at the keys ... stay tuned.

[Nameless Voice]

Some modifications to the NVSetQVarTrap and NVTrigQVar scripts which might help you. Check the documentation.

• Added r operator to NVTrapSetQVar
• Added m operator and NVTrigQVarAllowRepeats param to NVTrigQVar

[LarryG]

I shave my head and bow down in homage. Thank you.