Idea: How to take the haxx out of RNG

[TimerClock14]

i don't understand the example nor the suggestion.

from the discord example all i see is a doubled effect, instead of scrambling once it will scramble two times so in theory if you have only 20 time quantum you can end with only 20 earth quantum. I dont see how it can reduce the 'bad luck rng factor'. Suppose you play only with one element, you are still screwed this way and it will only advantage the rainbow decks by more evenly distribute the quantums.

how do you apply this to mutation for example? Mutate the creature once them mutate it another time but which mutation do you keep? how do you combine this? if you dont combine them it still like a single mutation.

same goes for FG matchups, you choose one FG then another and then ...? You pick the first or the second ? it still the same

This would subsequently change the game so that it would become less dependent on luck and more dependent on skill, as it should be.

A card that involve luck is part of the strategy in how to use/deal with it. If you want a game solely based on skill then remove all the luck cards (lots of shields, lots of entropy  cards, ...)
And i find it more important to resolve the mulligan first then to address the rng use. Using a proper mulligan like magic the gathering is the way to diminish the luck factor and improve the skill factor, but that's something the creator is opposed to and as such i dont see how your suggestion to diminish the luck can be approved.

Allow me to rephrase the example:

We start with a tray that has 12 slots, each slot represents one element in your opponent's quanta pool.

For the sake of the example, I'm going to make up some numbers here:
We have:
18
4
25

So 3 of our 12 slots are now containing quanta. The rest are empty.

(skipping to the end of a given turn in which discord successfully attacks the opponent)

We now must randomly take 10 quanta and randomly distribute them across all 12 slots (elements) in the opponent's quanta pool.

So we dump everything in the tray inside a large bowl.

We take 20 pieces at random from that bowl, and place them into a smaller container.
We take 10 of those 20 pieces, at random, from the container.

The remaining pieces are returned to the large bowl unaffected. And sorted back to their respective slots (elements) in the quanta pool.

Those 10 pieces (not returned to the bowl) will represent the quanta our discord will scramble.

Now we simply perform the inverse of the steps above to place the 10 quanta into random slots.

What was done here is simply perform a 2-part, 2-step randomization. Currently, (as of v1.27) the game only performs a 2-part, single-step randomization.

[catalyzeme]

We take 20 pieces at random from that bowl, and place them into a smaller container.
We take 10 of those 20 pieces, at random, from the container.

The remaining pieces are returned to the large bowl unaffected. And sorted back to their respective slots (elements) in the quanta pool.

Those 10 pieces (not returned to the bowl) will represent the quanta our discord will scramble.

Now we simply perform the inverse of the steps above to place the 10 quanta into random slots.

What was done here is simply perform a 2-part, 2-step randomization. Currently, (as of v1.27) the game only performs a 2-part, single-step randomization.

What you described would change literally nothing. The probabilities would be exactly the same. Are you sure you described what you are thinking properly?

[TimerClock14]

We take 20 pieces at random from that bowl, and place them into a smaller container.
We take 10 of those 20 pieces, at random, from the container.

The remaining pieces are returned to the large bowl unaffected. And sorted back to their respective slots (elements) in the quanta pool.

Those 10 pieces (not returned to the bowl) will represent the quanta our discord will scramble.

Now we simply perform the inverse of the steps above to place the 10 quanta into random slots.

What was done here is simply perform a 2-part, 2-step randomization. Currently, (as of v1.27) the game only performs a 2-part, single-step randomization.

What you described would change literally nothing. The probabilities would be exactly the same. Are you sure you described what you are thinking properly?

Yes.

[QuantumT]

TimerClock is correct when he says this would tend the results of RNG calculations towards the average. I'll show how with the simpler example of Chaos Power.

Right now we can think of Chaos Power as rolling a 5-sided die, with equal probabilities for each. So the distribution looks like this:

+1=1/5=20%
+2=1/5=20%
+3=1/5=20%
+4=1/5=20%
+5=1/5=20%

Now to do it Timer's way, we'll roll twice and take the average. First the chances of rolling the different number totals (pre-average):

+2=1/25
+3=2/25
+4=3/25
+5=4/25
+6=5/25
+7=4/25
+8=3/25
+9=2/25
+10=1/25

Now we'll divide by 2 to find the average. For the odd numbers I'll randomly choose whether to round up or down.

+1=1/25 + 1/25            = 2/25
+2=1/25 + 3/25 + 2/25 = 6/25
+3=2/25 + 5/25 + 2/25 = 9/25
+4=1/25 + 3/25 + 2/25 = 6/25
+5=1/25 + 1/25            = 2/25

Expressed in percents:

+1 = 8%
+2 = 24%
+3 = 36%
+4 = 24%
+5 = 8%

[Pineapple]

But I thought the purpose of "random" mechanics were to be...well, RANDOM.
We'll want to make everything less and less about chance and eventually we'll end up with something like "My rogue has 20% dodge, so if your warrior hits 4 times, his next attack must miss!"

[Terroking]

Another way to fix it, would be to insert a psuedo-random system. If none of you know what that is, here's an example.

Your opponent has 3 Graboids and 3 Shriekers out. You have a Dusk Mantle out.

If your opponent attacks, let's say the order is Shrieker->Graboid->Shrieker->Graboid->Graboid->Shrieker.

The Graboid attacks, and the RNG picks 1 or 2. If 2, it misses, if 1, it hits. Shrieker 1 hits, as does Graboid 1. Shrieker 2 misses, and then Graboid 2 hits. Now, because 50% of the opponent's creatures have hit, all others are guaranteed to miss.

So, basically it uses the % chance to put a limit on how many times you could succeed or fail.

For things like Discord it'd have to be on a per turn basis, e.g. is it takes 10 quanta the first turn, then it'll have to take 1, 2, 3, ect. up until 9 before it can restart.

I like Kakerlake's idea for deck shuffling though.

[Pineapple]

Most cards are balanced with the pretext that they are "random", and switching the RNG to "pseudo-random" would most likely unbalance many of those cards...

[Kjaska]

Expressed in percents:

+1 = 8%
+2 = 24%
+3 = 36%
+4 = 24%
+5 = 8%

This kind of RNG will lead to people relying on the 36% +3 much more and when the 8% +1 hits, they will rage even harder :3

Less RNG will no doubt give us more outcomes based on skill, but it will also expedite migration towards cookie cutter builds & make the game more boring. An Elements PvP match, although plagued by the current netcode issues from time to time, is still not very long. By playing a best of series instead of 1 match, one can reduce the impact of RNG without taking the pace out of the single duels.

[Dark Weaver]

TimerClock is correct when he says this would tend the results of RNG calculations towards the average. I'll show how with the simpler example of Chaos Power.

you misinterpreted what TimerClock is doing. And catalyzeme is correct saying that TimerClock's view changes nothing.
The thing you're doing is randomness with a non uniform distribution, and would just privilege some outcome over others. I dont see how you could apply this to hatch for example as there is no average between creatures...
If i take your example, timerclock idea is to roll the dice twice to have 2 results then flip a coin to choose which dice to apply.

RNG to "pseudo-random"

well the rng IS a pseudo random number generator
http://en.wikipedia.org/wiki/Pseudorandom_number_generator

The Graboid attacks, and the RNG picks 1 or 2. If 2, it misses, if 1, it hits. Shrieker 1 hits, as does Graboid 1. Shrieker 2 misses, and then Graboid 2 hits. Now, because 50% of the opponent's creatures have hit, all others are guaranteed to miss.

So, basically it uses the % chance to put a limit on how many times you could succeed or fail. 

In HommV the ghost can evade 50% of the attack i think, but there is a cap of 3 successfull misses. If you missed 3 times in a row, you are sure to hit next time you attack the ghost and the miss counter is reset to 0. I dont know any other game that alter the random behaviour in such a way, but its a good example on how to implement this kind of modification.
The problem with your solution is that the creatures attacking first have a better chance to  inflict damage than the last creatures. So in your scenario it's better to lay big attack creature first.

I like Kakerlake's idea for deck shuffling though.

i think its awfull :p, well not for the player but its too overpowered, you dramaticaly improve the efficiency of all decks. I think a proper way to deal with bad hand is to do a MTG mulligan system with loosing 1 card for each hand draw, so you're trading a card with the probability to get a better hand but you have to choose wisely.

[TimerClock14]

TimerClock is correct when he says this would tend the results of RNG calculations towards the average. I'll show how with the simpler example of Chaos Power.

you misinterpreted what TimerClock is doing. And catalyzeme is correct saying that TimerClock's view changes nothing.

You misunderstand. I responded yes to catalyzeme's question, not a rhetorical "yes" in response to his prior statements in that quote.
Allow me to clarify:
He asked me if I was sure I described what I was thinking properly. I answered yes.
QuantumT's explanation is much better than mine, however, and he's right on the money with it.

[Dark Weaver]

He asked me if I was sure I described what I was thinking properly. I answered yes.

So you didnt explain properly what you where thinking. (and the answer should have been no). The way you choose the 10 quantums in your example (with the bowl) is exactly the same way i choose the effect with the dice example.
And What QuantumT does only works (tend to average) because the effect is based on numbers.

Everything not mulligan that's currently decided by the RNG is affected by this suggestion

You cannot average things like choosing a creature for hatching eggs, choosing a god for next matchup etc...

[Kjaska]

If you appoint strength/usefulness rating onto the hatched creatures and false gods, i.e. a firefly is a 2 and a firefly queen is a 5 or Neptune is a 1 and Dream Catcher is a 10, you could utilize the method proposed. You would then get very easy and very strong FGs less frequently, than average FGs.