The ‘lucky’ Greg

DBJones

I think they did mess with the gem dropping formula somehow, I've seen more 5-matches for both my opponent and me than I'm used to. Well, ignoring my green ramp deck. I mean natural 5-matches are showing up a bit more for both sides. I think I still get as many now as when Greg would just let me have every one of them.

octal9

edit: n/m

Coilbox

octal9 said:

Coilbox said:

I think i've seen enough in the last few days to firmly believe that:

1. The gems that the table creates for you and the AI every turn is 'slightly' tweaked to give your opponent an advantage.

lol

wrong wrong wrong wrong and wrong.

Do you have something to prove me wrong? If you do, mind sharing it?

It doesnt matter how many times you say ‘wrong’, it wont give you any more credit.

theobserver

He did. Read the thread from the beginning. 

Coilbox

I did see what he shared about the RNG, that kind of scientific randomness mess on the first image might apply for training grounds. The other events? im not quite sure.

Im not basing my comments on any science or anything, only on what i can observe, and  it's a drastic difference on how any events work compared to training grounds.
The drop on victory rate given by bad gem matches and bad card draws is too high to call it just 'bad luck'. And it's systematic.

I might be wrong and then be the unluckiest player outside training grounds ever

octal9

Coilbox said:

Do you have something to prove me wrong? If you do, mind sharing it?

It doesnt matter how many times you say ‘wrong’, it wont give you any more credit.

I do have something to prove you wrong. In lieu of sharing it, as some are aware, I invoke the Fifth. Let's just say my curiosity knows few limits.

This exact question is what led me to independently verify the source of the official stance taken by the devs - the AI doesn't know what gems are coming in. 

It's something I've known and thus, argued for a long time.

FindingHeart8

Coilbox said:

I did see what he shared about the RNG, that kind of scientific randomness mess on the first image might apply for training grounds. The other events? im not quite sure.

Im not basing my comments on any science or anything, only on what i can observe, and  it's a drastic difference on how any events work compared to training grounds.
The drop on victory rate given by bad gem matches and bad card draws is too high to call it just 'bad luck'. And it's systematic.

I might be wrong and then be the unluckiest player outside training grounds ever

@theobserver

to be fair, while I applaud the attemps of octal9, volrak, etc to provide some sort of system of measurement for the player base, we don't have enough information to reach any conclusions beyond correlative theory.

Our client-side results (from what we pull from packs at least) is based on an honor-system; and we don't know if the coding template the designers use is the same or different than templates referenced in the articles listed.  If Hibernum developed their own template for gems, it would invalidate any related posts on how other gem-swapping games run.

@octal9

Who was your source again?  Didn't you make a thread about it once?

bk1234

I believe Octal has clearly answered these questions and provided proof of his statement, further tagging and hounding him does no good. 

Coilbox

octal9 said:

Coilbox said:

Do you have something to prove me wrong? If you do, mind sharing it?

It doesnt matter how many times you say ‘wrong’, it wont give you any more credit.

I do have something to prove you wrong. In lieu of sharing it, as some are aware, I invoke the Fifth. Let's just say my curiosity knows few limits.

This exact question is what led me to independently verify the source of the official stance taken by the devs - the AI doesn't know what gems are coming in. 

It's something I've known and thus, argued for a long time.

Look there is something im not denying. You defend all the time the fact that 'the AI doesn't know the incoming gems because they are not yet generated' and i can perfectly agree with you. Ok.
But the way the gems are generated independently of what the AI knows or not can still be balanced towards the oponent favoring his turn, his colors and whatever Greg might need for winning a battle. Do you have proof of this not happening?

Please show it to me, i cant wait to see it.

There are some battles in which no matter what you do, you know that the AI is going to win cause everything comes just in hand for him. He draws the perfect cards to answer whatever you do, he has the perfect gem match in the precise moment with a cascade even when in your turn there was NOTHING to match.
You play a support, he destroys it, you play a creature, he steals it or destroys it, etc, etc...
And, then again, it can happen in training grounds indeed, but how often it happens in some other events (specially pvp events) is just... far more often.

You seem to have information about the core code of the game and how it works, I would like to know what is your source to sentence all our suspicions (im not the only one at least who thinks the way I do).

Would you mind telling me where did you get this information from? 

Thanks in advance.

Coilbox

bken1234 said:

I believe Octal has clearly answered these questions and provided proof of his statement, further tagging and hounding him does no good. 

Im not... hounding anyone. We are just arguing about the subject. And... to be honest, i can't see unquestionable proof provided. He's free to answer or not, im not putting a knife in anyone's neck to get information.

James13

I think he's not sharing the underlying code for very good reason.

I'm willing to trust the word of someone far more knowledgable about the inner workings of the game.  And also willing to accept that it's just random and not programmed to be more complicated due to Occam's razor.

In my opinion, people are just noticing that the AI is better at picking relevant matches on the board now.

That said, I've already noticed a couple shortfalls and non-optimal choices even though it's overall strength is generally better.

Krishna

ElfNeedsFood said:

this happened as soon as they outlawed Smurfing tbh. and I couldnt be happier.

They outlawed smurfing in Platinum?  

Well, the Smurfing matchmaking exploit where you can pair your lvl 60 PW against a lvl 12 PW. People still be using low level PW

FindingHeart8

bken1234 said:

I believe Octal has clearly answered these questions and provided proof of his statement, further tagging and hounding him does no good. 

I'm not hounding him, but I looked through the links he provided and not finding this proof mentioned, which is what I'm inquiring upon.

Firinmahlazer

Maybe the word source being italicized in his previous post is some kind of code for something? I don't understand why people are so hell bent on thinking the AI is cheating. Maybe some people aren't as good as they thought when the AI actually puts up a fight.

Matthew

Krishna said:

There was definitely a dumbing down of the AI a couple months back, which surprisingly hasn't come up yet lol. It's probably a little better now than before that AI nerf. I miss the days when the the AI or Greg was 'The Computer' or CPU.

This was actually brought up, and has been consistently injected into conversation since Hibernum dialed Greg back several months ago.

The link cited by @octal9 really is a must-read. Objectivity is not an easy thing to practice, but it's absolutely necessary to truly understand what's actually going on.

Coilbox said:

THAT is what i mean. Im not crying or whinning or complaining about having a smarter oponent or a more difficult challenge as long as it doesnt involve tweaking the AI to just leave you with no good matches, THAT IS NOT FAIR PLAY.

I did stop using Koth too, cause its just ridiculous... the difference when you play with him and when you play against him is just... beyond noticeable.

You mean you don't strategically make off-color matches just to smother the AI and limit their mana gains? That's the implication I'm getting from this comment. Knowing when to take an off-color match instead of an on-color match is top-tier strategy, and it means the AI is actually more intelligent. I'm ecstatic that Greg is doing this now. It means he's actually playing intelligently, and it means I need to pay more attention. So I don't see how you can say you're complaining not about the intelligence level when that is exactly what you're actually talking about. Furthermore, if I can do it to Greg, why shouldn't he be able to do it to me?

As for Koth, I honestly can't remember the last time I used him in a non-TG event. He's far too limited in what he can do. Realistically, you're probably better off not using him.

Stormbringer0 said:

A good AI is fine, but not getting the cards you need while the opponent has literally an answer to any card you cast is plain unfair. I didn't lose matches because Greg is op, i lose matches becauce I draw the wrong cards at the wrong time - even with cycling and card draw.

I haven't noticed this at all. Are people actually complaining about this? I really have to think it's just more perception bias that is being directed at the wrong target, namely, the deck instead of Greg's beefed-up brain.

Coilbox said:

I did see what he shared about the RNG, that kind of scientific randomness mess on the first image might apply for training grounds. The other events? im not quite sure.

Im not basing my comments on any science or anything, only on what i can observe, and  it's a drastic difference on how any events work compared to training grounds.

I find it a little ridiculous that you're asking someone to back up his assertions, while at the same time you freely admit that you're not doing the same for your own, and saying that they are merely what you've observed. Furthermore, the information that he's already shared is about as scientific as you can get, and not some "kind of scientific randomness mess". So I don't know what more you could want from him on this one.

theobserver said:

He did. Read the thread from the beginning. 

Hey there [redacted], nice to have you back with us

[MOD NOTE] Post edited to removed player name used without permission [//MOD NOTE]

FindingHeart8

Firinmahlazer said:

Maybe the word source being italicized in his previous post is some kind of code for something? I don't understand why people are so hell bent on thinking the AI is cheating. Maybe some people aren't as good as they thought when the AI actually puts up a fight.

Everything posted in these threads has to stand up to the test of disagreement.

Overall I'm not on the boat of thinking the AI is cheating.  But I'm also not going to blindly accept thirdhand information as fact.

Octal has made a significant claim.  One I am interested in reading into.

Gilesclone

The best argument I can make against a cheating AI is that the best players are still getting perfect and near perfect scores.  Not once, but in every singe event.   If the AI was truly getting better gems, that would be impossible.

bk1234

Gilesclone said:

The best argument I can make against a cheating AI is that the best players are still getting perfect and near perfect scores.  Not once, but in every singe event.   If the AI was truly getting better gems, that would be impossible.

My HoD score was my highest ever this week -- the changes I noticed in Greg are not luck based, they are skill based -- ordering cards differently, choosing match 5's, etc. When you pay attention and predict his new behaviors you can build around him.

We had a great discussion in another community about this and I think those involved walked away with some great ideas on how to build for smarter Greg, perhaps this kind of discussion would help here as well. 

octal9

FindingHeart8 said:

Everything posted in these threads has to stand up to the test of disagreement.

Overall I'm not on the boat of thinking the AI is cheating.  But I'm also not going to blindly accept thirdhand information as fact.

Octal has made a significant claim.  One I am interested in reading into.

Without doubt, you shouldn't unquestioningly believe me. Its something you should be skeptical about.

I know a thing or two about the AI. Why? Because I'm insatiably curious and game AI/development has always intrigued me.

Curiosity is a good thing, unless you're a cat. I'm trying to not be a cat. Aren't you in ThePower9's discord?

Kinesia

One thing that people have said is that gem changers don't change loyalty gems anymore. (This was a huge hidden change that I'm not sure all the implications were thought out about.)

And when there are lands on the board I'm seeing more cascades for ME as well as Greg. I'm just seeing more overall, fullstop and bias makes it looks like they are all going to Greg but they aren't.

octal and Volrak are quite open to discussion but you should really be bringing them EVIDENCE instead of FEELINGS. Log your games honestly, record every cascade for yourself as well as the AI, record match 5s on both sides, record the match 5s you MISSED but noticed immediately after you did it even before the gems are falling.

And not just one game... You need logs of many many games to diminish the effect of the randomness.

Maybe you are just having a bad week and forget all the good weeks you had before?


Greg had his vision fixed, he no longer has that particular deliberate blindspot but he still has many other flaws to take advantage of.