Continuing worsening boards

Yeah… I agree.

The How:

I started counting reshuffles with mono last year. It was easier with the raid environment at the time.
And then was trying to recruit someone to raid rainbow and count reshuffles for me because… well…
I didn’t want to :rofl:

I imagined raiding with rainbow vs. gtv would be like swallowing my own vomit. There’s not enough Tums on the planet for me to put in the effort…
And wouldn’t you know… I couldn’t get anybody to do it… lol…

And then before I hit 600 samples I was really getting tired of playing mono… I’ve never enjoyed mono… it bores me to tears. Also it increases the risk. Also also, the best data for this argument is 3-2 data… that’s always bothered me too.

Anyway I quit mono at 632. And told myself to suck it up, and put together a rainbow attack squad. To my surprise…
I’ve enjoyed it!
I’ve posted a few shots of it around the forum. I think in the Fenrir thread. Showing off my Magni.


The Why:

My hypothesis is, that there is some form of weighted average which is meant to act as a pity safeguard. To help protect players from the possibility of extended / harmful RnG.

I don’t view millions of players as necessary for these events.
Although… that certainly won’t reduce frequency!
The extended timeframe that we play (years). We can all be exposed to a vengeful, unhappy RnGeesus. And such exposure would result in population loss…
“screw this game, I haven’t won in weeks. Bye” Which would affect income…
And why leave $ to chance when there is a solution? Let the seed include raid history!

This business strategy could work both ways.
As sort of a guarantee…when the ebb becomes flo
Utilizing social media as free advertising for firework displays and lucky pulls. Which works… when your alliance mate shows a single pull resulting in a new 5* + Hotm. Guess who’s next to drop gems…

And that this weighted average then causes waves of good and bad luck. Which if documented over time… appear on the graph as a healthy RnG. The ebb and flow go both ways…

So for me, how do you prove that? How do you disprove it?..

I figured well… if your due for a beat down with mono. The lack of one color means an abundance of another color or colors. And that means…

connect options!

So next… what’s a bad board with rainbow? Easy…
There’s only one 3-connect option.
Four 3-connects later. Your getting thwacked. Those are the same boards that tend to reshuffle!
So…
If there is an imbalance between the two?..


The Result:

I counted to 650 and sucked up some more mono which surprisingly pulled another one!

Rainbow reshuffled 19 / 650
Mono reshuffled 12 / 650

And… your right… this could still just be random…
It’s not enough samples.

And I’m having enough fun with rainbow. And can still see the ebb and flow go both ways. Also and…
Now it’s mono gameplay that gives me heartburn!

That I just don’t care about this topic anymore.
In fact…

I believe that this tl;dr will be my last here. The original topic that brought me here! Where I was immediately chastised by @Rigs :kissing_heart:

I just hate seeing some of the believers be frustrated or ganged up on by non-believers. A change in color stacking can relieve a good portion of pain that mono players experience. And choosing to play without pants on

Or… some other more intentional and mathematical approach

Can be fun too!
Anywho…
Good luck with your project @TheSchmoo

Chief Shill here. You called?

Sorry on late reply, been on a zoom call with the SGG Board of Directors hand picking next weeks victims for bad boards.

Maybe you’ve been lucky and missed it this time!

You do realise that you are agreeing to the opposing viewpoint, right?

That is plainly just ridiculous. The mods do an amazing job of dealing with the cubic tons’ of manure that is dished out daily by forum users. Whenever they post they do so referencing their large pools of experience and perspective and their commentary is always reasonable and logical. And they never enforce it on anyone - I have never seen a moderator make a post that amounts to “arguing”. Censorship is triggered by community flagging of posts, so the moderators basically follow through with what the forum is asking of them.

Your post sounds a bit shillish, to me.

I would love to see videos of 10 consecutive raids from you showing this…

Unfortunately I have to say that your assumption is wrong.
Reshuffling the boards is done for two reasons:

  1. There is no way to formed at least a 3-connect of any color.

  2. The board starts by itself, as 1 or more 3-connects have already been formed on the start board.

(3. I never saw that a color was represented under 2 tiles.)

The whole thing for all colors.

I can’t see the connection between good and bad boards.

Incidentally, these cases occur in a functioning RNG.
With manipulated boards it would be easy to prevent it.

Of course, one can argue about the quality of the RNG generator.

Your operating system’s RNG will definitely be used to create the maps.
Anything else would be too slow.

Summons are performed on the server side.
So sometimes you have to wait for the result.

I assume that both are not deterministic.

But I have to see that as sufficient.
It is impossible for your mobile phone to get deterministic RNGs without using a third party provider.
On the server, the effort would be considerable to generate it yourself.
Manufacturing requires a huge physical measuring facility.
In addition, the response times would be far too long.

See also:

Ergo the existing RNG must be sufficient, everything else would destroy the game.

That’s fine. You probably won’t.

Especially without proper definition in the first place.
Which is yet another matter of debate:

I contend that a bad MONO board has Many connect options. Which is usually supplemented from one color missing (the color you chose to attack with). And that causes an abundance of 1 or more other colors. Hence the many options of off-colored tiles you get to charge your enemies up for 1 point of damage each.

You can disagree and say that is a great board.
It won’t change my mind.

I further contend, that a bad RAINBOW board has only one connect option.
You can only connect 3 tiles out of 35… The other 92% of the board has no options.
After you send up your only 3 tiles, if the following tiles can’t make a match the board will reshuffle.

You can disagree and say that’s a great board.
It won’t change my mind.

So… I believe that if you already can not match 92% of the board. Well then… the probability is much higher that your about to experience a reshuffle.

And… if the algorithm is perfectly random. Then there is no weighted values for actual bad boards. Therefore… reshuffling should happen in equal amounts whether you use rainbow. Mono.
Or attack with 1 Aife. It’s all random. The algorithm doesn’t care what color you took. So there should NOT be a sizable difference.

However… if it is true.
And @Cheds is pushing the…
“Go Flick Your Own Face” button…
And he’s pushing it a lot. He does it in waves!
Well then… a discrepancy will develop.

You can disagree with that also.
I won’t argue. The @Cheds button was never a thing for me anyway!

The hard part is. And it was for me too.
Defining what a bad board is for rainbow…
People have plenty of mono data. And 3-2 data.

I wanted something a little more, outside the box
Something I could track. I’m just tired of doing it.
Data harvesting is fun for uh… someone? The game has given me plenty of extra work to do lately anyway

I disagree with your whole tangent, which I think is probably confusing both sides of the fence. But theres a vague notion that you are complaining about boards, so that side of the board manipulation fence will probably hesitantly rally behind it.

But you really don’t need to look at reshuffling as any kind of indication of manipulation of the boards. All you need to do is count the number of tiles in your majority color (mono, 4-1, 3-2, 2 -whatever), which has been done numerous times over numerous timeframes by numerous people, and see if you get close to the expected average. Which it has, every time.

What is the point in researching bad boards for rainbow? Who is complaining about comtinuing worsening boards for rainbow?

Not just rainbow.
Yeah that would be pointless. I need both, rainbow and mono. And I need a large sample, of each…

And if there is no discrepancy.

Then there’s no @Cheds button either

I don’t even want to argue.
I like discussions at eye level.
And I have to think about it again.

The hassle of simulating it in another software to check the condition is honestly too much for me at the moment to get independent data.

But maybe I’ll have an idea one day.

It’s really not hard to do. Count reshuffles.

The hard part is playing a tactic you don’t enjoy.

Simulating a generic RNG doesn’t reveal how this one was written.

If the sample is big enough, a couple years worth?..
:face_vomiting:

There shouldn’t be a large discrepancy. Anyway… I got 650 samples each. Took over a year. I have my results. I listed them. I don’t think someone read thoroughly. Not you @Barrista. I dig your appetite.
But I’m full.
You guys do your thing. Imma go do mine
with no pants on

First thoughts on generating comparative data from an independent source.

Creating a 5x7 array and filling it with RNG numbers from 1 to 5 is no problem.

Checking for 3 existing connections should also be easy.
(Here you can save existing 2-Connects.)
In this case there is no further check.

Then you can check a possible 3 connection at all 4 corners for 2 connections each.
Note these down for each color individually for further statistical evaluations.

Make sure there is at least every color and safe value.

Generate an output screen with the array for control.

Create a table of results for automatic runs.

After 100-200 manually initiated test runs, any number can be run automatically.

In addition, one could achieve results for the distribution of the individual colors.

Usually it is programmed quickly and dirty in 3-4 hours.

Unfortunately, I’m sick at the moment and don’t have the software at home to implement it quickly.

In addition, many would probably describe me as biased and see the data as manipulated as well.

Edit:
If you then automatically evaluate 100 or 1000 strips per 650, you can create a distribution curve.
This will result in a Gaussian curve.

Then you can assess the quality of the data collection and the data quality with the 650 interval.

Keyword 6sigma.

First thoughts are: “no problem”

Uh… I have a different first thought.

"How does any independent source
reflect this RnG?"

Are you using the same seed?

It makes me laugh to point at the discobot and say look see. There is proof you are crazy.
Proof?.. proved what exactly?

@discobot roll 35d5

I only have 20 dice. Shameful, I know!

:game_die: 4, 2, 5, 3, 1, 1, 5, 2, 2, 2, 1, 4, 3, 3, 1, 5, 1, 1, 4, 5

So this guy went through the trouble to make a video of his graph

And in this video he claims that he counted starting board colors. And out of 350 samples they were within a fraction of a percent! No deviation?!..
I ask you how is that random?..

image

Your RnG should also reflect this nearly perfect and equal supply of color. Or it doesn’t reflect this RnG

@discobot roll 20d5

:game_die: 2, 2, 5, 4, 1, 5, 1, 1, 1, 1, 4, 1, 1, 1, 1, 5, 1, 5, 2, 3

number color appeared percent
1 Red 18 30%
2 Blue 12 20%
3 Green 7 11.7%
4 Yellow 11 18.3%
5 purple 12 20%

@discobot roll 20d5

I just need help understanding how, this translates to a different RnG. Since they are all different.
And they are All Imperfect

1 Like

:game_die: 2, 4, 3, 4, 2, 5, 1, 2, 4, 2, 4, 5, 2, 3, 1, 4, 5, 5, 4, 3

The RNG only has the task of generating 35 numbers from 1 to 5.
Every RNG generator should be able to do this job sufficiently well.

The 20 samples from Diskobot are clearly too small.

With large samples like this 350x35, the match becomes calculable.

Your assumption is based on 650 starting boards.

That’s 650x35 RNGs.

The random numbers themselves are questioned by you as well as their position on the board.

If these are often created independently of one another under the known conditions, it becomes clear to what extent the results are reproducible.

And you will see if your values ​​are in the range that is likely for a working RNG.

If so, your numbers prove that the RNG and the distribution on the board are what they should be.

Ah. So if you can produce similar results. Then we can extend the sample?

I think I understand. This way I don’t have to play mono for another year.

Awesome man! Thank you! I hope it works :+1:

If someone tells you a random number from 1-5 20 times, you will never get such an answer.

But if you take a dice and roll the dice 20 times (6 does not count = reroll), you will definitely get such stripes with the same numbers.

But it doesn’t have to be, so you have to repeat it enough times to get a large enough sample.

Then you will come to such results.

Cookie Settings