Manipulating RNG in Battles - Predicting "Luck"

[magickzzl]

Well seeing a Scizor use Protect via Wi-Fi means the person is extremely noob, or excellent at RNG abuse. Taking measures to abuse it should be punished (and will again, try to be by WI-Fi'ers). Knowledge of how to abuse RNG is fine, but it will be obvious that you are abusing it if you're running a logical Scizor set that aims to make use of protect at a random time.[/QUOTE]

It could also mean that they use that scizor in doubles with an exploder, and hence adds in an unpredictable factor. You can end the game and call them a cheat possibly for not.

Again, research is great, and really, do you think that prediction is effective in WIFI? really? if you do you need to get on there more and fight some fly articunos.

 

[CCM30]

I'm still not following how you can tell. There is no possible way for you to know what someone is thinking. Someone can easily be either

A)RNG abusing
B) Just playing normally
C) Really ****ing lucky

How on earth can you tell? You can't. There is no way to tell intention, stream of consciousness, purpose without being the person yourself.

Plus, the ability to really compute such an action would be ridiculous without a computer or a hacking device. It's an interesting thought, but I can't see it viable.

 

[cantab]

Plus, the ability to really compute such an action would be ridiculous without a computer or a hacking device. It's an interesting thought, but I can't see it viable.

I said this before, but actually, I'm sure SOMEONE could do the RNG in their head. (From what I've heard Euler could probably have.)

 

[Coldcut]

I'm a bit surprised no one has drawn a parallel to counting cards in Black Jack. In both cases, "hidden" information is used to gain an advantage. My personal opinion is that, like counting cards in casinos, RNG manipulating should be banned from all forms of competitive play. However, as with counting cards, it is very difficult to prove. My intuition tells me that allowing RNG manipulation to go unchecked would polarize the metagame into those who can/do and those who can't/don't, since, over time, those who "predict luck" will rise above those who don't.

As a person with a heavily scientific mindset, I would be opposed to a ban on all research of the subject, since knowledge, when used correctly, is beneficial. However, like with counting cards, I don't think this is what the "spirit of the game" is meant to be, and should be kept out of competitive play. A separate server/ladder would be a perfect place to test this, though...

Just my thoughts.

 

[imapedarsag]

yes, someone probably can do it in their head. but how do they know what seed they hit?

edit: i too believe that this should be researched, but placed in a seperate server/ladder.

 

[Little Green Yoda]

It certainly adds a new wrinkle to the gameplay but the difficulty will likely turn many people off. Now imagine if one player's actions influenced how the RNG of the second player's game worked in link battle!

I said this before, but actually, I'm sure SOMEONE could do the RNG in their head. (From what I've heard Euler could probably have.)

Calculations aside, I don't see how that would be possible without knowing all the spreads on a given seed. It would take a tremendous amount of research to figure all that out.

 

[Paradox]

I think the idea wasn't to have people start abusing this on wi-fi or in game, or even in typical shoddy battles. That would be pretty much impossible, unless you are constantly inputting things into a sophisticated computer program alongside the battle.
I think (and correct me if I'm wrong) that OmegaDonuts' suggestion was to include this information in shoddy battles (presumably not in standard battles, but perhaps in a separate server or as an optional clause) to add a new level to the game without actually changing any game mechanics. It also is, for now, pretty much entirely theoretical, the amount of work required to calculate effects and re-program shoddy to act in the same way as the cartridges (RNG wise) would be immense.

I for one think it is a very interesting concept. It would be really cool to have to think "Well, if he bullet punches he'll get a crit and KO, or he could swords dance because I'll obviously see that and protect, or he'll protect, seeing as how my quick claw will activate this turn and I could be packing will-o-wisp." It would be adding a new layer to the game without changing mechanics, which is a very interesting concept.

 

[ΩDonut]

What I wanted to see was speculation on what a game would look like if both players knew in advance what the state of the RNG would be, and both players could adjust their tactics accordingly. What moves would you make if you were playing such a game? Would people load up their teams with Protect and Substitute? (Probably) Would they add hax items like Quick Claw to mess with the opponent's RNG prediction, resulting in the rise of Banette with Frisk?

You're free to discuss how such a discovery would "ruin Pokemon forever", but if you do you're really missing the point of this thread. At least, my point anyway.

Also, reprogramming Shoddy to do this would be relatively trivial. Just use a pre-defined "random" sequence and have it use that for all its random number purposes. It could also show players what a move's results would be, for a certain series of numbers, because all those calculations are a bit much for mental math.

 

[AJC]

I see one big potental negative impact the ablity for a defensive booster to negate the chance of critcal hits the normal thing that stops such boosts from being a game winner.

By abusing the RNG to never have him/her be critted he which would make stuff like calm mind suicune alot more annoying to face.

 

[cantab]

Shell/Battle Armour might become popular. Serene Gracers would nosedive, but might not become worthless - indeed, they'd probably be the only Pokemon who could still make SOME use of added effects, since there's still a chance of the secondary effect activating twice in a row and protect not blocking it the second time.

Also, it's not just crits - being able to predict whether a damage roll would be low or high might count. You can be certain your Pokemon will marginally survive, and leave it in, whereas normally you'd be taking a gamble.

I think the higher power, lower accuracy moves would decrease in usage, because knowing your opponents best attack against you would miss gives you a cheap chance for setup.

Though an interesting question on accuracy - how exactly does it work. In particular, if, for example, Fire Blast will miss, will Stone Edge always also miss (if the RNG simply rolls a value and checks it against the move accuracy), or is there a chance Stone Edge would hit (if things are more complicated)?

In the unlikely event damage rolls are off a different RNG to effects rolls, every player wants at least some moves that roll for effect. Otherwise the opponent can keep a successful effect roll 'in hand', waiting for the opportune moment to unleash for example a paralysing, FP on same turn, Thunderbolt.

 

[Zog]

One thing I'd really like to know is... how are you supposed to work out where the RNG is during a battle?

To my knowledge there's no feasible means of making sure you know where the RNG's at mid-battle. Sure, you can get a specific seed you want beforehand, but that takes time and to be honest I'd rather catch an easy shiny than have an idea what will happen in the battle. And even then, it probably won't work because you don't know what the opponent will be using. One <100 accuracy move or other event involving the RNG could potentially screw up all your plans.

Really, this isn't going to work. While it is interesting to think about, it just isn't realistic.

 

[Intro1827]

Say you get to know the seed mid-battle. From then, you can track any seed advances and react accordingly. You only need to know how many times a signle move/action requires the RNG, and if you're faster, you can already know if a move is going to hit, if it's going to critical, have a secondary effect, and so on. You could even predict your foe using, say, Focus Blast, and try and advance the seed just enough for it to miss. Of course, if you're slower, it becomes tough

 

[Andy Snype]

RNG Abuse in battle is a very interesting concept, to say the least. Predicting crits and damage variation is quite crazy, but how would someone find the seed? It could vary from Game cartridge to game cartridge, trainer ID/SIDs of both players which is near impossible for two random wifi battlers to know the other person's trainer ID/OT/SID. The delay between the players' choices and wifi connections would also do it. Hell, even the clock and date could vary from DS/DS Lite/DSi's, likely causing different seeds. If someone wants to try it, go for it.

I doubt that one person can learn all of this info, calculate the seed, and still time the moves right for "unlimited hax." Therefore, this research isn't harmful. Although it could help us understand how the game literally functions.

On a side note

Again, research is great, and really, do you think that prediction is effective in WIFI? really? if you do you need to get on there more and fight some fly articunos.

Don't battle on the serebii chatbox then.

 

[magickzzl]

On a side note


Don't battle on the serebii chatbox then.

Serebii has a chatbox? I'll have to try it! Ive been randomly getting challenges via livejournal. lol. But I was using that as an expample of you never know what your opponant is going to use on the wifi. You can try and predict but there are some strange things that go on, even from experianced battlers. Shoddy is far more predictable.

But yes great points about the DS times, id's etc. My DS time is way off and Im just to lazy to fix it, and i know im not the only one. It is intreging though, and as said before, there will be SOMEONE who will be able to figure it out in there head, infact people may already be doing it subconsiosly. When you put as many hours into a game as some people do around here you get into the rhythems.

 

[cantab]

How to find where the RNG is - Action Replay. No need to modify anything, just a code to show it somehow. This is of course cheating, but so subtle it would be hard to detect over WiFi.

Or in a long battle you might be able to discover it by sufficient observation of behaviour. Not gonna be doable by a human, but plug into a computer every crit, miss, flinch, secondary effect, and especially exact damage values, and it should find it. As a (rough, not really correct) estimate, a "will-it-or-won't-it" event gives one bit of information (several attacks will do multiple tests), and damage figures give I think 5 bits (especially if you make a guess as to the opponent's stats, reasonable if you know they use RNGed Pokes or legal hacks). The RNG seed is I think a 128 bit number. So I reckon 10 to 20 turns to establish where the RNG is. Quicker if you specifically set out to find it, for example by using elemental fangs a lot (though they're rubbish, so you probably don't want to).

 

[imapedarsag]

You know what's really wierd? Ever since RBY, I would use confuse ray and when it was their turn to move, I would press A and B in a random, patternless order, making sure that there was clearly no pattern and I swear that every time I did it, the opponent hit himself in confusion. Just this morning I killed someone in the battle factory by doing this with a dusclops. Also, another "trick" that's worked for me is, when I'm confused, I keep pressing A, each push spaced out evenly, starting from when the birds spin to when they end, and again, I swear I hit through confusion everytime. I always press some sort of button combination when I'm in a battle. When I'm plain out attacking, I just hold A and it goes on the higher side of the damage spectrum. When I'm hit, I hold B and it goes on the lower side of the damage spectrum.

Maybe I'm just really lucky in-game and it just so happens that the RNG wants it that way. But I really don't think so. The RNG may control Critical Hits, but there is no way that I have been "lucky" for 12 years. Just thought I should share my in-game schemes.

 

[cantab]

That may just be confirmation bias. Kind of luck the old Up+B business.

Do each of these things about 20 times, on any game you like, keeping track of how the effects go. If you get a statistically significant difference, then there may be something in what you are saying.

 

[Cathy]

This thread is mostly relevant for WiFi friendlies only.

In a competitive tournament, a judge would handle setting up the games so that you have no control over the initial seed, and it would be impossible to calculate the seed, or predict the next number, from the limited amount of data you get during a match. As such, any battle simulator will simulate this tournament setting, so you would not be able to predict random events in a simulator either.

That said, pokemon where you can predict when effects will occur is an interesting parlour game, perhaps worth of some analysis in its own right. I just mention this since already people were commenting on the implications for simulators.

 

[WECAMEASROMANS]

Yes, this method would control the amont of "luck" in a battle. But it would completly fuck up the metagame. Knowing when something is about to occur is just too much, and battling competitively won't even be fun anymore. Plus, luck is one of the metagame's largest factors. You can't simply just decide to cancel it (or use it to your advantage) just like that.

This seems interesting, but I don't think we should go through with it.

 

[Showsni]

Surely if player A and B are battling over wifi player A's RNG will handle any of A's calls to the RNG and player B's will handle any of player B's, and then report the outcome to the other player? So, whilst you might be able to tell where your own RNG was, you'd have no information about your opponent's, unless you could manage to work out the seed simply by watching what moves they do and the outcome.

 

[Mario With Lasers]

Hmm, I think you all are forgetting something. How the hell would you be able to know the seed of your opponent? Both players will take a part on the battle, even if you know the seed you entered the battle with, you cannot know the opponent's with reverse engineering (well, not without a few turns, I guess).

This won't be something really practical to do. Just chill down, and focus on the main topic of the thread: the luck-controlled metagame.

You know what's really wierd? Ever since RBY, I would use confuse ray and when it was their turn to move, I would press A and B in a random, patternless order, making sure that there was clearly no pattern and I swear that every time I did it, the opponent hit himself in confusion. Just this morning I killed someone in the battle factory by doing this with a dusclops. Also, another "trick" that's worked for me is, when I'm confused, I keep pressing A, each push spaced out evenly, starting from when the birds spin to when they end, and again, I swear I hit through confusion everytime. I always press some sort of button combination when I'm in a battle. When I'm plain out attacking, I just hold A and it goes on the higher side of the damage spectrum. When I'm hit, I hold B and it goes on the lower side of the damage spectrum.

Maybe I'm just really lucky in-game and it just so happens that the RNG wants it that way. But I really don't think so. The RNG may control Critical Hits, but there is no way that I have been "lucky" for 12 years. Just thought I should share my in-game schemes.

Something like that worked in Yellow. Repeatedly pressing A in a certain speed made all of Brock's moves miss my Pikachu. I almost defeated his Geodude and Onix only with Quick Attack (but then my sister bothered me and... well. I lost the timing).

 

[pikachu25]

You know what's really wierd? Ever since RBY, I would use confuse ray and when it was their turn to move, I would press A and B in a random, patternless order, making sure that there was clearly no pattern and I swear that every time I did it, the opponent hit himself in confusion. Just this morning I killed someone in the battle factory by doing this with a dusclops. Also, another "trick" that's worked for me is, when I'm confused, I keep pressing A, each push spaced out evenly, starting from when the birds spin to when they end, and again, I swear I hit through confusion everytime. I always press some sort of button combination when I'm in a battle. When I'm plain out attacking, I just hold A and it goes on the higher side of the damage spectrum. When I'm hit, I hold B and it goes on the lower side of the damage spectrum.

Maybe I'm just really lucky in-game and it just so happens that the RNG wants it that way. But I really don't think so. The RNG may control Critical Hits, but there is no way that I have been "lucky" for 12 years. Just thought I should share my in-game schemes.

I myself proved that pressing A and B on a GBA with Red version plugged in right when a pokeball opened made it a 100% catch rate if you got it right. That's how I got all of my Abra's lol...

Anyway, for something on-topic. I think that a metagame in which both people knew what the results of the next, say, 20 (or more, depending on how it would be implemented online), RNG calls would add strategy to the game. I for one would find this to add prediction to the metagame, but it should be an option or another ladder/server.

Also, let me know how I can help with research. This kind of thing is right up my ally :).

 

[DaNukes]

WOW this is really an amazing idea.It would certainly revolutionize the game like many people have said, but i definitely think pokemon would become MUCH different.

For example, now, we basically are removing all luck and replacing it with skill/prediction. Do you guys think this will ruin the purpose of 100% accurate moves (well i guess the opponent can't mess your RNG up if you use it)?

I think this will definitely make lax incesnse and brite powder incredibly useful, because your opponent will never know when to factor in those items and when not to, and it also makes the reliance on 100% accurate moves moot becuase the opponent could just be exploiting it to make you miss all the time.


HOWEVER, all this speculation is based on teh fact that you can determine the RNG based on a list, but I am pretty sure that it would gamefreak would design the game so that the call to the RNG and the ingame effects are based on the time the move is selected or something similar, becuase if it isn't then it can be abused in the way you have just said and their game that has been sold worldwide will be completely broken

I think the most important thing that this will do (if its possible) is make the game MUCH more confusing and complex, because now there is a level of predicting something else, how the oppoenent is manipulating the RNG, so it is no longer just the opponent's moves and switching, it is that AND how they are manipulating the RNG for that move, and how you need to counter both, or just one to be safe.

 

[Zeke50100]

If you had photographic memory and an inhuman sense of time, RNG "abuse" would be an advantage (I put it in quotes because I consider it more as an act of taking advantage of the RNG rather than abusing it).

Anything with the current state of the RNG is possible, but the feasibility? Meh. The metagame would be dominated by teams that rely on perfect RNG calculation and prediction, and those will probably see very little use in "standard, non-RNG abuse" play.

Of course, what I said is all based on the assumption that the RNG in these games are time-based, which is commonly seen in RNGs of other games. If it's merely affected by a sequence of events, and not the time in which those events took place, knock yourselves out.

 

[petrie911]

I don't know how you'd possibly abuse the RNG on an actual cartridge. The RNG draws a number every frame during battles, so combine that with Wi-Fi lag and you have something completely impossible.

However, that's not what the TC is talking about. A Shoddy mode where the RN list is visible to both players would be fascinating. Manipulating the list to try to get criticals where you need them while your opponent is doing the same...it would add a new level to the game. It kind of reminds me of my experience with the GBA Fire Emblems. They used the same RNG sequence for the map every time, so you could deduce the approximate sequence of numbers by experiment, and manipulate things to go your way.