Judgement Mechanics Study

Jihnsius · Apr 20, 2007

The other thread was locked (probably because of all the flaming in it,) so here's another official one dedicated to the study of the Judgement Hammer. There are three main areas in this study:
1. Patterns. Is there some sort of pattern to what appears to be random?
2. Probabilities. If there aren't patterns, or we don't know them yet, what are the probabilities of each number showing up on either a large or specific scale?
3. Gameplay. Is there some sort of way to play G&W so that the Judgement is in your favor (in the most common case, getting a 9 most of the time?)

We know it's not random, due to these four rules:
1. The first hammer in a match cannot be a 1 or 2.
2. The second hammer in a match cannot be a 2.
3. Any number the hammer displays cannot be obtained again until after 2 more hammers.
4. A death (fall or SD) will reset the rules currently applied to #3.

This doesn't rule out, however, the possibility that the number obtained after following these rules isn't random.

- 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.

Probabilities known of so far:
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.

EDIT: This thread has served it's purpose. With no new, useful data coming in to prove or disprove anything, I motion that this thread be locked.

Umpy · Apr 20, 2007

very nice but i have gotten a 1 on the first hammer of the match and it happened today i promise you

Supreme Pie Ninja · Apr 20, 2007

Not to be off topic, but for when KevinM reads this, stop. Stop typing your oh-so witty comments in these topics, and just walk away.

KevinM · Apr 20, 2007

Supreme Pie Ninja said:
Not to be off topic, but for when KevinM reads this, stop. Stop typing your oh-so witty comments in these topics, and just walk away.

Not to be on topic but lets look at the reasons why even in the other thread i continued to tell you guys to stop trying its random..

1. Multiple people in the other thread would say that you had to die to reset it or something along those lines

My friend got either three or four i believe it was four without it dying

2. The fact that you were basing a lot of your equations and patterns off the fact that you can't get a 1 or a 2 off the first throw and the man above you just said he did..

Look i'm not trying to be a kill joy but there is not a single good argument here that isn't broken down into the fact that its random, just like Turnip Pulling and Misfires.. they're random and can't be controlled

Is that better for you if i try and state my facts that way or just because i have a differing opinion that seems pretty plausible you don't like it?

Jihnsius · Apr 20, 2007

That is more reasonable, and I respect that you have an opinion, but the fact that we, as a group, have probably done at least 10,000 tests on this and all of them hold up to the rules we've found pretty much negilfies your claims. If you want to stay in this thread and have an active attempt to -disprove- us, that's more than welcome, but flaming as you were doing in the last thread is intolerable.

As for multiple numbers in a row: If it's as possible as you claim it to be, find out the methodology behind how you did it, because any known set method of tests prove otherwise.

EDIT: And another note on methodology, if it does turn out that one person has done what we have speculated is impossible would, in fact, prove that there is methodology to it, and that we're just doing it incorrectly, making this thread all the more valid.

KevinM · Apr 20, 2007

Here's how my friends four nines went the other day

Match begins.. he hasn't used his hammer yet and neither of us has died.. he's at i believe 60 percent and me as Roy (don't ask i did a random character) is at 83 percent.. he whips it out and sunnuvagun the first time he uses it is a nine..

Fast Forward
He hasn't died yet (have i mention i suck with roy :chuckle:

and he has been just kinda taking potshots with the Manhole at me and staying away i'm on my third stock at around 45 percent and i have him to 93 percent.. Pulls out the hammer.. boom thats another nine.. he hasn't used it yet and its only like maybe thirty seconds later

This is the worst of it..
I'm at zero percent and he is talking about how the nines being in his favor today and he laughs and yells Nine and hits me with yet another nine sending me outwards out of FD and yet i come back.. and he's hysterically laughing and yells nine again jumps out and nines me to death within i'd say 3 seconds...

If thats not random then i give up disproving this hah

Jihnsius · Apr 20, 2007

What version was this on? We're yet to do tests concerning the game timer, and the combination of methods, but if what you say is true, it'd have to be a combination of these methods, and not random.

KevinM · Apr 20, 2007

I got it on launch day and still have that disk soo.. 00 i guess?

Jihnsius · Apr 20, 2007

That may be a good area of study, but I don't know anyone that has 00.

KevinM · Apr 20, 2007

How can i check the version to be sure?

Jihnsius · Apr 20, 2007

I don't think there's any surefire indicators other than bugs and glitches. Easiest one would be to pick Bowser, jump, and right when you hit the ground, press B. If the flame comes out instantly (flame canceling,) it's 00.

Supreme Pie Ninja · Apr 20, 2007

Alright, I appreciate that you actually took the time to explain yourself Kevin, instead of being all smart about it ;D

But, I think that these topics aren't exactly trying to prove that it's not random, rather that there are some other things to it. Like the old "can't be same number twice in a row" (example, as this seems to be in the process of being busted flat) can apply to this, while the numbers themselves are still random.

Conditions, that's what I meant. We know it's random, we're just trying to find conditions that go along with it.

And on the topic of your crazy friend there...I'm not saying I don't believe you, but whenever you're using/playing G&W, try to record the matches in any way possible, just so that if it happens again, we'll have concrete proof, as there are many people out there that are harder to fool than me xD

EDIT: I have my GC hooked up at the moment in the next room, so I'll be testing more with it:

5 stock, 1v1, Final Destination, 5 Hammers pulled, suicide, repeat

(just to sort of mesh the current theories together)

Pasqual · Apr 20, 2007

I may take some time this weekend and look through all the vids in the G&W video thread, and tabulate the hammers and whatnot. The only issue with that is I won't be able to tell if there are early-frame-interrupted hammers, which is one of the current theories.

My coding project is at school and if I can resolve the last couple errors I have, should be done Monday.

Jihnsius · Apr 20, 2007

Supreme Pie Ninja said:
We know it's random, we're just trying to find conditions that go along with it.

We don't know it's random, we just assume as much, because that's how it appears to be. The only way to be 100% sure that it's random is if someone from HAL/Nintendo announced that it is, otherwise there's no method to determine if a sequence of numbers is random or not, we can only determine if it's not random (and until that's determined, we'd have to determine it as inconclusive.)

Regardless, the mission is to determine how the Judgement Hammer works.

KevinM · Apr 21, 2007

Well i'd be glad to record some matches.. just take a moment of silence cause my gamecube got busted yesterday... its the most heartbreaking thing to ever happen to me.. so when i get another one, (their only like what 50 bucks used now) i'll invite my lucky friend over

Umpy · Apr 21, 2007

the version i did the 1 on the first one was also 00 if you wanted to know that anyway

Johnknight1 · Apr 21, 2007

Intesting study and finding. I thought it wasn't all random, and I noticed some random patterns, but whenever I played at my friends house it was always random. Then I found out he had version 00, and I had another version=wierd. I love how the versions of the game can be so diffrent, even with these little random things.

thebluedeath1000 · Apr 25, 2007

I was playing as jigglypuff against game and watch in a friendly, he leaped and hit me with judgement hammer nine, I made a joke about him not being able to get another, he leaps, I shield, he gets another nine...I thought that wasn't possible, I keep my shield up, he leaps and does the hammer again..and guess what, it was a nine, it shatters my shield and sends me into the sky.

Jihnsius · Apr 25, 2007

thebluedeath1000 said:
I was playing as jigglypuff against game and watch in a friendly, he leaped and hit me with judgement hammer nine, I made a joke about him not being able to get another, he leaps, I shield, he gets another nine...I thought that wasn't possible, I keep my shield up, he leaps and does the hammer again..and guess what, it was a nine, it shatters my shield and sends me into the sky.

Must've been version 00. From now on, I'll be sure to never play anything other than 02.

Needle of Juntah · Apr 26, 2007

good GW's know when its time.

Needle of Juntah · Apr 26, 2007

you didnt say judgement cant happen 2 times ina row *applause*(even though it 10X rarer than pulling 2 broke turnips in a row)
also judgement likes to come out first in a match, AT THE END of the match.
3rdly never rely on a judement, because there really is no way to remember all the patterns in the game. if you ever did i would laugh at you because GW is so crappy, you should probably learn how to play fox or sheik or falco perfectly with the amount of time it would take to learn over 100 patterns the game has for -> B
hell even i dont know all those patterns, but i do know some

Jihnsius · Apr 26, 2007

Needle of Juntah said:
hell even i dont know all those patterns, but i do know some

Really? Tell me, then.

thebluedeath1000 · Apr 26, 2007

lol..And I really have no idea on the version so you're correct more than likely.

Jihnsius · Apr 26, 2007

Finally. 10,000 hammers (unique from previous experiments), Kongo Jungle vs Falco, standing still at the spawn point. 1153 nines. 1/8.67

It took me 4 different sessions to complete, so that probably had a slight effect on the outcome, but it should be fairly rounded. I was expecting to see it a lot closer to 1/9. I'll brainstorm the possibilities later, time to sleep.

HyugaRicdeau · Apr 26, 2007

Frankly I don't believe anything that anyone claims contrary to the known laws (the 2 previous numbers cannot be the next one, 1 and 2 start in the queue, death resets the queue, otherwise the number is selected from the 7 possible at random) without raw unedited video, or a statistical study that demonstrates something significantly different than random, and neither should anyone else.

If you (not you specifically Jihnsius, but anyone in particular) want to do a scientific study of the GW hammers, you must uphold it to the standards of science; to anyone that claims something like they got 4 9's in a row without dying (although nothing prevents you from getting 4 9's on one stock, so long as the 9 has been out of queue), the burden of proof is on the claimant. Man up or shut up.

I would also suggest you research binomial distribution so that you can understand the meaning of statistical significance in this case, although in the large number of trials you are using, the Poisson distribution is good enough and much easier to deal with.
http://en.wikipedia.org/wiki/Binomial_distribution

The only question I really have about the hammer is when exactly does it count as the number having appeared? That is, if you hit fwd b but you get hit before the hitbox comes out, does that count as getting the number you WOULD HAVE gotten?

Jihnsius · Apr 26, 2007

HyugaRicdeau said:
The only question I really have about the hammer is when exactly does it count as the number having appeared? That is, if you hit fwd b but you get hit before the hitbox comes out, does that count as getting the number you WOULD HAVE gotten?

Already looked into that and yes, it does.

HyugaRicdeau · Apr 27, 2007

Jihnsius said:
Already looked into that and yes, it does.

Cool thanks. Can I ask how you figured that out? I thought of recording 2 trials, then hitting GW out of fwd B before the number comes out, and seeing if I ever got a number from the previous 2 trials, as that would mean they had left the queue.

Jihnsius · Apr 27, 2007

HyugaRicdeau said:
Cool thanks. Can I ask how you figured that out? I thought of recording 2 trials, then hitting GW out of fwd B before the number comes out, and seeing if I ever got a number from the previous 2 trials, as that would mean they had left the queue.

Bingo. I used Fox and shined G&W out of the hammer as soon as I could (not frame perfected, but close.) Did this a few hundred times and never got the same number within 2.

DH_Ninja · Apr 28, 2007

i like it and believe the mechanics of 1 out of 7 and never same number twice. i used this in my game and watch game *who i was thinking of using as my secondary, which would be very unusual since i main as zelda. anyway, ive been being carfeul with the judgement and watching in the first four. *which is where ive been using them . Anyway, ive been getting many nines lately in my matches. matter of fact, a few minutes ago when i was playing my friend, i used the judgement three times and believed i was going to get a nine the fourth time. So i put my self into a posistion where i could hit with the judgement, i hit with the nine and won the match.. anyway, i don't think the percentages have much to do with the hit of the judgement though.

Rapid_Assassin · Apr 28, 2007

DH_Ninja said:
i like it and believe the mechanics of 1 out of 7 and never same number twice. i used this in my game and watch game *who i was thinking of using as my secondary, which would be very unusual since i main as zelda. anyway, ive been being carfeul with the judgement and watching in the first four. *which is where ive been using them . Anyway, ive been getting many nines lately in my matches. matter of fact, a few minutes ago when i was playing my friend, i used the judgement three times and believed i was going to get a nine the fourth time. So i put my self into a posistion where i could hit with the judgement, i hit with the nine and won the match.. anyway, i don't think the percentages have much to do with the hit of the judgement though.

Lol why would that be unusual? *cough, cough*

The hammer is probably random, other than the fact that you can't get 9's twice in a row without dying. And by random, i mean one of those pseudo random programs, because computers can't be 100% random.

HyugaRicdeau · Apr 28, 2007

Yeah, I agree that it is "probably" random (or pseudo-random); of course we can't prove that it doesn't depend on some strange obscure factor in the game without actually looking at the code.

But I don't think it's unreasonable to posit that depending on something like % or stage or character is "minimally likely," that is to say, you might as well be searching for the Celestial Teapot.

Jihnsius · Apr 28, 2007

Considerable trial and error and number crunching raises some interesting results, and aside from that everything is just trying to break down the logic of how things work.

This came to me in a dream just now: In order to make a countdown timer for matches, it seems much more likely that the timer resets every match rather than being based off of some "global" timer, or the system clock on your NGC. If the JH is based off of any in-game timer, I'm sure we could figure out which one it's based on with a little bit of persistence. When a match starts, the controller input starts being functional at "GO!," and I'd imagine that most players' reaction times are great enough to be able to input a forward B within 5 frames of it starting. If so, and if the JH runs on a counter that's reset every match, we could figure it out by starting a new match and JHing right off the bat and then compiling the set of numbers you get. Even if you rarely go up to 10 frames before you input the forward B, the median will be around 1-6 frames, which should give a rounded 6 different JH outcomes.

Johnknight1 · Apr 28, 2007

Jihnsius said:
Considerable trial and error and number crunching raises some interesting results, and aside from that everything is just trying to break down the logic of how things work.

This came to me in a dream just now: In order to make a countdown timer for matches, it seems much more likely that the timer resets every match rather than being based off of some "global" timer, or the system clock on your NGC. If the JH is based off of any in-game timer, I'm sure we could figure out which one it's based on with a little bit of persistence. When a match starts, the controller input starts being functional at "GO!," and I'd imagine that most players' reaction times are great enough to be able to input a forward B within 5 frames of it starting. If so, and if the JH runs on a counter that's reset every match, we could figure it out by starting a new match and JHing right off the bat and then compiling the set of numbers you get. Even if you rarely go up to 10 frames before you input the forward B, the median will be around 1-6 frames, which should give a rounded 6 different JH outcomes.

Intriging thought about this whole issue. I have no idea about how the whole thing works, but it is definitely intresting, to say the least.

alphameric · Apr 29, 2007

Has anyone checked to see if the hammers are sequential?
So like, each number lasts a couple frames.

So every second it runs through 7 of the numbers in sequential order.
So, if smash is 60fps then every 8.5 frames the next number is called up.

Jihnsius · Apr 29, 2007

I think I've pretty much confirmed it: It goes by the match timer (not a global timer or NGC clock). Here's the list of numbers I got upon forward Bing the first instant:

9 9 9 4 5 9 4 4 4 6 4. I quit after that because I thought it was obvious.

This doesn't rule out, however, that there are methods to change what you get, but it's a step to figuring it out.

HyugaRicdeau · Apr 29, 2007

Sorry to be a killjoy, but that doesn't really prove anything. You got a bunch of 9's and 4's in a statistically small sample. I tried to reproduce these results using the method you describe and I didn't get anything different than usual. Also, The match timer doesn't actually start until a few split seconds after "Go!" even though you CAN move around during that short interval. It is enough time to Fwd B in fact, so even if the number were based on a match timer, it's not the one at the top of the screen.

Jihnsius · Apr 29, 2007

HyugaRicdeau said:
Sorry to be a killjoy, but that doesn't really prove anything. You got a bunch of 9's and 4's in a statistically small sample. I tried to reproduce these results using the method you describe and I didn't get anything different than usual. Also, The match timer doesn't actually start until a few split seconds after "Go!" even though you CAN move around during that short interval. It is enough time to Fwd B in fact, so even if the number were based on a match timer, it's not the one at the top of the screen.

I noticed that, too, that the actual match timer didn't have an effect (since it doesn't start until after "GO!" disappears, which is enough time to do a forward B and them some, which is even more noticeable when you're trying to do BTT or HRC,) but perhaps it's the actual counter that starts as soon as the match starts, considering I got a substantial amount of 4s and 9s, which may very well be perceived as one millisecond after the next. I'd say if you want to try to help out (and I can pretty much assume you have a perfect understanding as to what I'm trying to do, as you seem to be a pretty intelligent individual,) just keep a log of each number that you get to try to add to the database that we have so far. Just note, however, that this is purely constructive and I don't want any *******s lying about their achievements just to try contradict me.

GimR · Apr 30, 2007

Why doesn't someone with Action replay do it frame by frame and use forward B on each millisecond and each hundreth of a second to see if thats how it works.

Anyone got an action replay?

alphameric · Apr 30, 2007

To clarify, my theory is that the judgment hammer is sequential.
Some of the other rules still apply:
1. 1 & 2 Can't be drawn first
2. Drawing a number X will require using the hammer two more times to possibly get X again.

Sequential means that every 8.5 frames the 'counter' runs though 7 of the 9 numbers in sequential order.

The numbers do not need to start from the lowest number.
IE: Round starts and the counter starts at 7, as opposed to 3.

The numbers might go sequentially UP or DOWN (only one of them, no idea which), meaning the count can go 1234 etc; or 4321 etc.

The numbers might loops, or count up, then down again.
123456789 -> 1234566789 or 12345678 -> 9 -> 87654321

But thats generally my theory.
If someone has AR and could test it (speed hammering through frames?), that would be awesome.

Jihnsius · Apr 30, 2007

Makes sense.

The only problem with the match timer theory is that 100ms doesn't transfer over into 60fps too nicely, 40 of those milliseconds won't show up.

Judgement Mechanics Study

Smash Lord

Smash Apprentice

Smash Ace

TB12 TB12 TB12

Smash Lord

TB12 TB12 TB12

Smash Lord

TB12 TB12 TB12

Smash Lord

TB12 TB12 TB12

Smash Lord

Smash Ace

Smash Journeyman

Smash Lord

TB12 TB12 TB12

Smash Apprentice

Upward and Forward, Positive and Persistent

Smash Champion

Smash Lord

Smash Champion

Smash Champion

Smash Lord

Smash Champion

Smash Lord

Baller/Shot-caller

Smash Lord

Baller/Shot-caller

Smash Lord

Smash Ace

Smash Master

Baller/Shot-caller

Smash Lord

Upward and Forward, Positive and Persistent

Smash Ace

Smash Lord

Baller/Shot-caller

Smash Lord

GimR, Co-Founder of VGBootCamp

Smash Ace

Smash Lord

Information

Network