TGM
Smash Lord
dude, there is not a link. this game has been out for 6 years. the chances of someone finding it out NOW aint gonna happen if they havnt found it already. face it. the 9 is random.
-
Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!
You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!
Controling the 9?
- Thread starter M3tr01D
- Start date
- Status
DarthLuigi36
Smash Cadet
- Joined
- Apr 12, 2007
- Messages
- 63
- Location
- Whittier, CA, USA
- NNID
- CmonBlueFalcon
- 3DS FC
- 3823-8676-0627
I think it may also correlate to the number of opponents. I was messing around in training mode, just randomly using Judgment and other attacks, G&W vs. three Marios. Often when they were at high damages, I'd pull 9s on them. It could be entirely coincidence, I wasn't keeping track (I should have been).
Jihnsius
Smash Lord
Yeah, when I'm practicing I usually use hammers where they normally shouldn't be now, just to see what happens. Can't figure anything out yet...
HyugaRicdeau
Baller/Shot-caller
What Quetzlcoatl had posted earlier is true. In case you missed it, I will summarize:
The number is completely random with the following constraints:
The last 2 numbers you rolled can't be the next one.
The game starts out with 1 and 2 in the queue.
The queue resets to 1 and 2 whenever GW dies.
Now, I don't have the program code to back this up, but this model fits every test I have done. I have never gotten 1 or 2 on the first hammer, and I have never gotten either of the previous 2 numbers on the 3rd, save for when GW dies in between. This doesn't mean that this model is exactly correct for sure, but AFAIK it's the only one that explains the data. I encourage everyone else to test this if they don't believe it.
And I have version 00.
The number is completely random with the following constraints:
The last 2 numbers you rolled can't be the next one.
The game starts out with 1 and 2 in the queue.
The queue resets to 1 and 2 whenever GW dies.
Now, I don't have the program code to back this up, but this model fits every test I have done. I have never gotten 1 or 2 on the first hammer, and I have never gotten either of the previous 2 numbers on the 3rd, save for when GW dies in between. This doesn't mean that this model is exactly correct for sure, but AFAIK it's the only one that explains the data. I encourage everyone else to test this if they don't believe it.
And I have version 00.
Jihnsius
Smash Lord
It may be true to an extent, but that's no denying that there are claims of getting the same number in a row. In fact, I think I got a 0 at one point in time. It was probably an 8 and I saw it wrong, but I'm sticking by the idea that it was a 0. Even if these rules are true, there still may be a way to control it within those parameters. We'll never know if we don't try.
HyugaRicdeau
Baller/Shot-caller
Haha 0 would be funny. But considering that lower numbers tend to be less beneficial for GW (like 1), I think 0 would be like 9'ing himself.
And claims contrary to what I've posted are really only just that. All my tests, and likely those that Quetzl did before me, I know I certainly wasn't the first to do them, are done in a controlled manner, and aren't "what happened one time I was playing my friend." Of course if there is in fact is some kind of esoteric factor like your %, or whatever, that affects the outcome, it will be really difficult to detect.
And claims contrary to what I've posted are really only just that. All my tests, and likely those that Quetzl did before me, I know I certainly wasn't the first to do them, are done in a controlled manner, and aren't "what happened one time I was playing my friend." Of course if there is in fact is some kind of esoteric factor like your %, or whatever, that affects the outcome, it will be really difficult to detect.
GimR
GimR, Co-Founder of VGBootCamp
Hey guys I think I might have found some sort of link, I only tested 100...
When you get a 5 than a 3, or a 3 than a 5, you will get a 9 50% of the time after that. I think I've come up with other things too, but I want you guys to test this out first.
I have version 1.2
When you get a 5 than a 3, or a 3 than a 5, you will get a 9 50% of the time after that. I think I've come up with other things too, but I want you guys to test this out first.
I have version 1.2
Jihnsius
Smash Lord
Kick ***. Keep it up. I'll test that in a bit.
GreenMarth
Smash Apprentice
)Jihnsius
Smash Lord
I was thinking: Maybe it goes in a pattern (such as when G&W is first unlocked) following the digits of Pi or something like that. That'd be neat.
why is this thread still open? lol its random, the end. who runs these boards?
Eggz
Smash Hero
Soooo the game and watch boards are the most ******** ones on SWF? Jesus.
Jihnsius
Smash Lord
In order for our "results" to be accurate to testing the idea that you can "set up" the nine, all reports for any given set will have to be the same method each time you use judgement hammer.
Here's a little something I wanted to try out, scan over these sets of numbers and see if you notice anything special:
On FD, opposite side of stage standing towards, versus Yoshi: 4375679471857346281256125961392542164134257263
On FD, opposite side of stage standing towards, versus Zelda/Sheik:
579459356432564265932896492386183459825431437568275674618742794574123
On PS, vs Marth, hitting him:
8563593825691893425785285674275287962713879162958128762753456249534923985429854374639278238652469723493428961596157143683243193748549149538219465196785
EDIT: We're talking about a theory here, guys. Don't be a douche and dismiss all theories as false just because you want to look cool in front of the kids. If you don't want to be involved in contributing to proving or disproving the theory, then stay out of the thread, simple as that. We've done nothing wrong to have this thread locked.
EDIT 2: Looking over the numbers, it kinda goes along with the theory that any given number can only appear again after two more hammers to reset the counter for that number. If the random number generator can't omit numbers to be generated (that'd take a bit more coding than just using if-else for non-omissions), that means that every time you hit the same number in the RNG within three hits in a row (9, 8, 9,) the second 9 would have to be changed. If this happened, I doubt they'd have it cycle through the RNG again, because that'd add another cycle to the processor, which could add up to visible lag in some strange cases, so the only logical way to do it would to be to use an algorithm to change the 9 into something else. Perhaps they would go back so many numbers in the list of previous numbers to another, or perhaps they'd run it through an equation to make it something else, etc, but no matter how they do it, if this is the case, this would be the first step to determining what's coming next.
If that paragraph was too confusing, let me try to break it down:
Do three hammers in a row. If the numbers are, in fact, randomized, as everyone seems to want to believe, then it'd be a lot easier to do this:
- Choose a random number 1 through 9.
- If it's the same number as one of the previous, change it to something else without running this loop again.
Than to do this:
- If the last number was 1, choose a random number 2 through 9.
- If the last number was 2, choose a random number 2 through 9. If it's 2, subtract 1.
etc. This isn't the only method to do it, this is just an example.
And this wouldn't work, because it'd add more frames to the animation:
- Generate 1 through 9.
- If it's the same number as one of the previous, go to previous line and start over.
Of course, this is hypothetical, and it would be possible to make it to omit random numbers from the generation in the first place, it'd just take more work to do than the simple "generate, throw out if already used."
EDIT 3: We can pretty much eliminate the idea that the numbers are derived in a pattern format, since every time you hard reset without saving, you'll get a different set of numbers. This shifts the idea of "controlling the 9" over to the idea of "manipulating the frequency at which certain numbers are randomly chosen." Although I don't think we should completely throw out the pattern idea, because we're not seeing every scenario here. I assume the general concept of what we've been hoping for is judgement hammer is random UNLESS a certain requirement is met, which would then put it into a pattern format (whether it be all 9s or a combination of numbers.) If this is the case, there'll have to be a lot more testing than just standing still, jumping, or anything done so far.
Here's a little something I wanted to try out, scan over these sets of numbers and see if you notice anything special:
On FD, opposite side of stage standing towards, versus Yoshi: 4375679471857346281256125961392542164134257263
On FD, opposite side of stage standing towards, versus Zelda/Sheik:
579459356432564265932896492386183459825431437568275674618742794574123
On PS, vs Marth, hitting him:
8563593825691893425785285674275287962713879162958128762753456249534923985429854374639278238652469723493428961596157143683243193748549149538219465196785
EDIT: We're talking about a theory here, guys. Don't be a douche and dismiss all theories as false just because you want to look cool in front of the kids. If you don't want to be involved in contributing to proving or disproving the theory, then stay out of the thread, simple as that. We've done nothing wrong to have this thread locked.
EDIT 2: Looking over the numbers, it kinda goes along with the theory that any given number can only appear again after two more hammers to reset the counter for that number. If the random number generator can't omit numbers to be generated (that'd take a bit more coding than just using if-else for non-omissions), that means that every time you hit the same number in the RNG within three hits in a row (9, 8, 9,) the second 9 would have to be changed. If this happened, I doubt they'd have it cycle through the RNG again, because that'd add another cycle to the processor, which could add up to visible lag in some strange cases, so the only logical way to do it would to be to use an algorithm to change the 9 into something else. Perhaps they would go back so many numbers in the list of previous numbers to another, or perhaps they'd run it through an equation to make it something else, etc, but no matter how they do it, if this is the case, this would be the first step to determining what's coming next.
If that paragraph was too confusing, let me try to break it down:
Do three hammers in a row. If the numbers are, in fact, randomized, as everyone seems to want to believe, then it'd be a lot easier to do this:
- Choose a random number 1 through 9.
- If it's the same number as one of the previous, change it to something else without running this loop again.
Than to do this:
- If the last number was 1, choose a random number 2 through 9.
- If the last number was 2, choose a random number 2 through 9. If it's 2, subtract 1.
etc. This isn't the only method to do it, this is just an example.
And this wouldn't work, because it'd add more frames to the animation:
- Generate 1 through 9.
- If it's the same number as one of the previous, go to previous line and start over.
Of course, this is hypothetical, and it would be possible to make it to omit random numbers from the generation in the first place, it'd just take more work to do than the simple "generate, throw out if already used."
EDIT 3: We can pretty much eliminate the idea that the numbers are derived in a pattern format, since every time you hard reset without saving, you'll get a different set of numbers. This shifts the idea of "controlling the 9" over to the idea of "manipulating the frequency at which certain numbers are randomly chosen." Although I don't think we should completely throw out the pattern idea, because we're not seeing every scenario here. I assume the general concept of what we've been hoping for is judgement hammer is random UNLESS a certain requirement is met, which would then put it into a pattern format (whether it be all 9s or a combination of numbers.) If this is the case, there'll have to be a lot more testing than just standing still, jumping, or anything done so far.
Pasqual
Smash Journeyman
Jihnsius, I think recursive calls could also handle the omission problem. There would be 10 arrays, one with all numbers 1-9 (call this All), than a Without1, Without2, etc.
Something along the lines of (in Java, at least) this, assuming from the first hammer throw of the match or whatever.
Int x = hammerthrow(All); //RNG's with the All array.
for(int y = 1; y < 10; y++
{
if(x == y)
--{
//some method that would set the next hammerthrow() call to use a Without(whatever x is) array
--}
}
The only issues I can see with that are possible LIFO troubles and the fact that, supposing it would have to be keeping out 2 numbers each time (one in its first iteration of being exlcuded, the other in its second), would require a good deal of arrays or some other method I can't think of right now.
Also, and I'm sorry that I've not read the thread thoroughly yet, but has anyone talked about the reapportionment of probabilities after the first two hammer throws, when there would only be 7 possible numbers each throw?
Something along the lines of (in Java, at least) this, assuming from the first hammer throw of the match or whatever.
Int x = hammerthrow(All); //RNG's with the All array.
for(int y = 1; y < 10; y++
{
if(x == y)
--{
//some method that would set the next hammerthrow() call to use a Without(whatever x is) array
--}
}
The only issues I can see with that are possible LIFO troubles and the fact that, supposing it would have to be keeping out 2 numbers each time (one in its first iteration of being exlcuded, the other in its second), would require a good deal of arrays or some other method I can't think of right now.
Also, and I'm sorry that I've not read the thread thoroughly yet, but has anyone talked about the reapportionment of probabilities after the first two hammer throws, when there would only be 7 possible numbers each throw?
t0m0
Smash Apprentice
No point in even having this thread... the nine is random.
Ensee
Smash Ace
Pasqual
Smash Journeyman
Hm, so looking back over the thread and at Quetzalcoatl's post, I see some of the stuff has already been taken into account. If we were to go on the assumption that the only rules are no 1s or 2s on the first throw, no 2s on the second throw, and the two-more reset rule, I think I could write up a Hammer class that would fit the theory well. Is anyone interested? I'd rather not do it if nobody cares since it'd take a bit of time.
DUDE I FIGURED OUT THE WAY TO GET THE NINE TO WORK EVERY SINGLE TIME
Its random GG
Its random GG
Jihnsius
Smash Lord
All contributions are appreciated. I'm having fun putting work into this.
That's what I'm saying, it'd take a lot more work than just doing another randomization and changing the number if it's already come up. Also, because it's an attack, it'd be defined in a script and not hard-coded into the engine, which would probably make it even less likely for them to put too much work into making it completely random. On top of that, the fact that we already know there are rules to the hammer is enough basis to say it's not random, and that there is a deal of order to it, whether small or large.
1. You can not start with a 1 or 2.
2. The second throw cannot be a 2.
2. If any given number is used, the earliest you can see that number again is after 2 more throws.
1/7 chance to see a nine on first throw.
1/7 chance to see it on the second throw if it was not on the first.
1/7 chance to see it on the third throw if it was not on the first or second.
1/7 chance to see it on the fourth throw if it was not on the second or third.
0 chance to see it on any throw if it's been seen in the past two throws.
These are the probabilities we know of thus far.
Wanted to check on that idea that higher percentages gives more nines.
FoD vs Kirby - 612 throws, 75 nines = 1/8.16
FD vs Y-Link - 600 throws, 72 nines = 1/8.33
And for low percentages:
FoD vs Kirby, downtilt to sh judgement - 503 throws, 58 nines = 1/8.67
FD vs Y-Link - 500 throws, 53 nines = 1/9.43
Also figured I'd check probability of nines as the first throw after each reset.
75 throws, 11 nines = 1/6.82
Check out the ratio for downtilt to sh as compared to standing still. I don't want to go on a limb and say that it's proof that there's a method, but it's still an awesome coincidence.
Pasqual
Smash Journeyman
I'd have to call that coincidence until a larger pool of data is accumulated, say in the 5-10 thousand area...daunting, to say the least.
I've got a preliminary version of a Hammer class running, should be done this week. I was thinking of also tabulating numbers from that but it'd be pointless in regards to Smash as I'm moreso doing this as code theory and practice.
I'm also wondering, what with the cases of people claiming multiple consecutive 9s without losing a stock, if there's also a timer reset variable. It'd be nice to see some data reflecting hammers under the same circumstances except for the time in between each one (preferably, once every 5 seconds/10 seconds/20 seconds/30 seconds/1 minute). There's also the possibility that the timer doesn't work as a preset countdown, but instead an exponential variable; being that, the chance of a number repeating starts off infinitesimal and exponentially increasing as the timer winds down. Those are purely theoretical, of course, but nevertheless possibilities to consider.
I've got a preliminary version of a Hammer class running, should be done this week. I was thinking of also tabulating numbers from that but it'd be pointless in regards to Smash as I'm moreso doing this as code theory and practice.
I'm also wondering, what with the cases of people claiming multiple consecutive 9s without losing a stock, if there's also a timer reset variable. It'd be nice to see some data reflecting hammers under the same circumstances except for the time in between each one (preferably, once every 5 seconds/10 seconds/20 seconds/30 seconds/1 minute). There's also the possibility that the timer doesn't work as a preset countdown, but instead an exponential variable; being that, the chance of a number repeating starts off infinitesimal and exponentially increasing as the timer winds down. Those are purely theoretical, of course, but nevertheless possibilities to consider.
Jihnsius
Smash Lord
I'll do some experimenting with the game timer and lengths of time.
Omg I Figured Out That The Hammer Is Random Do I Win Yet?
DarkKnight077
Smash Lord
I think..it's time.
Nappa:"What is the power level of Game and Watch Hammer?"
Vegeta:"It's over 9000!!"
http://www.youtube.com/watch?v=Bd2kFtJHp0s
Here ya go.
Nappa:"What is the power level of Game and Watch Hammer?"
Vegeta:"It's over 9000!!"
http://www.youtube.com/watch?v=Bd2kFtJHp0s
Here ya go.
Jihnsius
Smash Lord
Just get out of the thread.
Aiko
Smash Champion
Whenever I've got a spare moment, I'll filter hammers and look for the 9. I take a mental note of the 2 previous numbers and continue filtering till I get the previous number or both numbers in a row. It does appear that the 9 is more likely to appear after these numbers than others.
If the 9 appears after a totally unrelated number, then I also remember the predecessing numbers. It appears to me that after 3 or 4 of these different predecessors, that the next 9 will have the same previous numbers as the first pattern I memorised. I have actually said "next one is a 9" to my friends on several occasions and more often than not i'm right. So there will be some pattern, despite being obscure or difficult to work out.
If the 9 appears after a totally unrelated number, then I also remember the predecessing numbers. It appears to me that after 3 or 4 of these different predecessors, that the next 9 will have the same previous numbers as the first pattern I memorised. I have actually said "next one is a 9" to my friends on several occasions and more often than not i'm right. So there will be some pattern, despite being obscure or difficult to work out.
Jihnsius
Smash Lord
I think it's pretty much ruled out that there's a pattern, but if you'd like to try to see for yourself, go ahead. I'd recommend writing a list of the numbers you get so you can look over them as a whole to try to find a pattern. If you want, as well, I have a short list of numbers I got in a row a page back or so.
Dude my friend got four without dying...
DOES THIS MEAN ITS RANDOM?
DOES THIS MEAN ITS RANDOM?
Jihnsius
Smash Lord
That would be very meaningful, I'll give you that, but the fact that you're just talking out of your ass to increase your post count makes me respect you a lot less than you're given around Sacramento.
EDIT: Wait a sec, you're not Kev$. Nevermind, your opinion is nil.
EDIT 2: Another thing to note, so long as the forward B animation starts, it will count as the next number. In other words: If you input forward B on frame 1, and are hit on frame 2, it'll still count as the number coming out. If we put some research and time into this using frame perfection, I'm pretty sure we can either verify or rule out recursive randomization calls without omissions.
- Status