File → Movie → Uncheck Bind Savestates to Movies and Automatically Backup Movies
Config → Enable → Uncheck all settings
Movie goals:
Aims for fastest time
Manipulates luck
About this game:
Vegas Stakes is a casino gambling simulation game released for the SNES in the early 90’s. Players are given a starting amount $1,000 and given the goal of turning it into $10,000,000 through the use of betting on various games. These game include blackjack, poker, craps, roulette, and slots. There are several different casinos to choose from, each with a different limit on the betting amounts. Initially the player only has 4 casinos to choose from, each of which places a limit on how large of a wager can be made. After earning $100,000, the game will unlock the 5th and final casino which will allow the player to place bets without limiting the wager amount.
Improvements over previous run:
The introduction movie is now completed skipped. This was achieved by pressing the “X” button at the title screen instead of the standard “A” button. This was an odd inclusion on the developer’s part, because all other menus and dialogue in the game are cleared through the use of the “A” button.
The previous version of this run used the craps game to quickly achieve $100,000 within 2 rolls. This new run gets the money by manipulating the game into giving the player a winning lottery ticket worth $100,000. This new strategy saves time in a couple different ways. Much of the time is saved because the player no longer needs to switch the higher limit casino. As the lottery ticket can be given at any of the casinos, I can simply stay with the default low limit casino. This eliminates the long drawn out dialogue boxes and screen transitions. Additionally this lottery ticket method also saves time because it removes the betting and rolling animations that would be required for craps game. The lottery ticket can actually be given during any of the games, but blackjack was chosen for 2 unique reasons. With blackjack, you can opt out of hand and trigger a “random encounter” which in any other game would require some sort of dice rolling/slot/card dealing animation. Another benefit of choosing blackjack is how the RNG operates. Blackjack allows for complete control over 2 unique values which must perfectly align in order to trigger a specific event. Knowing these 2 values is how I managed to keep luck manipulation frames down to a minimum.
The final section of the run is almost the same as the previous version, whereby the player visits the no-limit casino and bets on craps to make the $10,000,000. I managed to keep luck manipulation frames down to a bare minimum here (~3 frames).
About random encounters and luck manipulation:
Throughout the game, there are several different “random encounters” that can occur with other casino patrons. You can visit http://www.gamefaqs.com/snes/588828-vegas-stakes/faqs for a detailed breakdown of all the random encounters that can occur. For the purpose of this run, I was interested in making the lottery ticket random encounter occur. The lottery ticket random encounter can yield the player a winning lottery ticket worth anywhere from $10,000 to $100,000. If I could get the lottery ticket worth $100,000, then I would be able to skip right to the no-limit casino without having to place any bets at all on the low limit casinos.
In order to get the lottery encounter to occur, I first needed to find out all the possible times that the lottery encounter can occur. I found 2 unique memory addresses, that when combined would determine the type of random encounter. These 2 memory addresses are 0x7E000B and 0x7E000A. Each of these values has 256 possible values. When you multiply the two values together, you get a total of 65,536 values. All told, there are 65,536 different possibilities when it comes to determining a random encounter. Only 25% of the values actually yields a random encounter, while 75% of the values will hit a blank. Out of those 25%, I had to narrow down to just the lottery ticket encounters and ignore all the other random encounters. To do this, I created a painfully large excel spreadsheet and began testing each individual value and putting them onto the spreadsheet. After a few hours, a very complex but repeatable pattern started to emerge. I will attach the spreadsheet below. When all was said and done, I had a total of 1,344 values which would yield the lottery random encounter. In order to quickly test all 1,344 values, I created a lua script which would pause the game when the 2 values lined up creating one of the 1,344 known lottery values. Each time the game would pause, I would make a note of the frame it occurred on and test it to see if it would be the $100,000 lottery ticket. If the ticket paid out less than $100,000, I would skip it and move on to the next paused frame. The values change every frame randomly during the deck shuffling sequence in the blackjack game. This allowed for pin point accuracy because I could let the sequence run for as long as I wanted and simply press A when I hit the desirable frame. On top of all that, I also had even more control over the randomness, because a seed is planted when you click on blackjack at the game select screen. This seed will determine how the 2 random numbers (0x7E000B and 0x7E000A) will play out. There are a total of 256 starting seeds. In all the other casino games, the values are fixed as soon as you click on the game you are choosing to play from the game selection screen, and only change once the current roll/hand is over.
Unfortunately it desyncs on my end.
The ROM is correct, as stated by BizHawk. I tried various versions of BizHawk. Only 1.1.0 interim and 1.1.0a skip the intro. After that it appears to desync (there are major delays and buttons appear to be pressed on input display when they have no effect). I've gotten as far as frame 7000 or so and all it shows is a blackjack table with nothing happening.
hero of the day, are you using any special options or firmwares?
I tend to have trouble video dumping in 1.1.0 since it commits a memory leak, especially if Capture OSD is checked. This problem is probably not in more recent versions of BizHawk.
Joined: 11/14/2005
Posts: 1057
Location: United States
FractalFusion,
I think I found out why it is desyncing. Can you please test out the run again with the following setting disabled:
Under file/movie > disable bind save states to movie and automatically backup movies
Under config/enable > disable all settings
I found that doing this allowed me to sync the run with the emulator after deleting my config file.
For those who can't get it to sync, I posted the run on youtube temporarily:
http://www.youtube.com/watch?v=VzrfKgHN67k
Got it to sync, that movie unexpectedly runs at full fps for me, newer versions of BH never do that.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
FractalFusion,
I think I found out why it is desyncing. Can you please test out the run again with the following setting disabled:
Under file/movie > disable bind save states to movie and automatically backup movies
Under config/enable > disable all settings
Thanks, hero of the day. It syncs now. It also solves the memory leak issues as well.
Very impressed by your run. I wish to know how you did this. Yes vote.
feos wrote:
Got it to sync, that movie unexpectedly runs at full fps for me, newer versions of BH never do that.
For me it's just BizSlug 1.4.1 that's the problem. BizSNES 1.4.0 runs at least as fast as BizSNES 1.1.0.
Maybe this is more of a general gambling question than a TAS question, but here goes.
You enter Laurel Palace with $101K. You bet $100K on a single roll of 3 (pays 15 to 1), meaning you get your original bet back plus 15 times $100K=$1.5M. Thus, your bankroll becomes $1,601,000.
Next, you play $1.2M on a horn bet. From what I understand, this is the same as betting $300K on 2, 3, 11, and 12 at the same time. When you roll a 12, you thus lose $300K on each of the 2, 3, and 11 bets, and gain 30*$300K=$9M from the 12 bet. The net gain would thus be $8.1M, which shouldn't be enough to reach $9,999,999 and win the game. Is the game's implementation of the horn bet wrong by paying 30 to 4 (or 15 to 2) upon your boxcars roll instead of 27 to 4, or am I misunderstanding how casinos work?
Edit: And what happens if you bet on a horn bet and roll 3 or 11? Do you get paid 15 to 4, or 12 to 4? If the bet pays 30 to 4 on 2 or 12 (2/36 of the time) and 15 to 4 on 3 or 11 (4/36 of the time), then this bet has no house edge whatsoever. o_O
When you roll a 12, you thus lose $300K on each of the 2, 3, and 11 bets, and gain 30*$300K=$9M from the 12 bet. The net gain would thus be $8.1M, which shouldn't be enough to reach $9,999,999 and win the game. Is the game's implementation of the horn bet wrong by paying 30 to 4 (or 15 to 2) upon your boxcars roll instead of 27 to 4, or am I misunderstanding how casinos work?
The true odds of rolling a 12 are 36:1. The house pays 30:1. But I do think the game makes an error. It calculates the winnings correctly (30*300,000=9,000,000) but I think it mistakenly returns the entire original bet (1,200,000) instead of 300,000. Thus the player's total money reaches $10,601,000 instead of $9,701,000.
Movie description needs to be updated, rather than a copy/paste of the obsoleted.
This movie, in fact, does not win money in a limits casino by rolling dice, but rather a generously given ticket.
Also, I have only one complaint about the run itself: It does not use an exclamation point for the final message. Other than that, short, sweet run showing what intense luck can bring in minutes, and a good improvement over the prior run. Not much else for me to say.
I have no clue about how casino works, so I'm not the one to fix it.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
When you roll a 12, you thus lose $300K on each of the 2, 3, and 11 bets, and gain 30*$300K=$9M from the 12 bet. The net gain would thus be $8.1M, which shouldn't be enough to reach $9,999,999 and win the game. Is the game's implementation of the horn bet wrong by paying 30 to 4 (or 15 to 2) upon your boxcars roll instead of 27 to 4, or am I misunderstanding how casinos work?
The true odds of rolling a 12 are 36:135:1. The house pays 30:1. But I do think the game makes an error. It calculates the winnings correctly (30*300,000=9,000,000) but I think it mistakenly returns the entire original bet (1,200,000) instead of 300,000. Thus the player's total money reaches $10,601,000 instead of $9,701,000.
Joined: 11/14/2005
Posts: 1057
Location: United States
Great description. Thanks for the quick publish guys.
I am still writing up my notes on the run. I will edit my submission text when that is complete with a breakdown of how luck manipulation and event manipulation works in this game.
I think this is the wrong place to request a new run of a different category.
I want to echo hero of the day and say great description to those who worked on it.
hero of the day,
I suppose you are very busy with writing up everything. I'd like to know one thing at least. What is the probability of getting the first event (lottery ticket to $100000) to occur right away?
Joined: 11/14/2005
Posts: 1057
Location: United States
I don't have access to my computer right now, so I can't give an exact number. I have a spreadsheet which lays out all the probabilities of getting the lottery event to occur. Getting the amount is a different beast all on its own. I simply tested every lottery event until I hit the 100k payout. I know that that payouts can vary from 10k all the way up to 100k, at intervals of 10k iirc. I'll try to get the sheet up soon, which will better explanation things.
