DEMA: Destroy 'EM All! V2.1

[jmdt]

So is this deck faster than the Ultimate Speed EM deck?

As far as ttw, yes its faster.  I do not however have conclusive stats as to if its a better score/electrum grinder than the Ultimate Speed EM deck.

[tinkady]

I been testing this on my account and Christmas on trainer...and I think christmas is not much slower while EMing almost every time and having more consistent draws. somebody wanna prove me wrong?

[tinkady]

on a side note, I just had a game where I accumulated 134 quanta without playing anything but a vamp dagger but was still one turn away from winning
and the next game was a beautiful 4 turn EM. shows how inconsistent it is lol

[jmdt]

I been testing this on my account and Christmas on trainer...and I think christmas is not much slower while EMing almost every time and having more consistent draws. somebody wanna prove me wrong?

Isn't it nice that we now have choices for EM decks

[tinkady]

Isn't it nice that we now have choices for EM decks

yeah but we still wanna find the best one. what i'm saying is that while DEMA is pretty great it doesnt EM enough.  It doesnt have enough generation to run more than 3 heals, which is not enough a lot of the time.
how goes testing on christmas (or santa? whats the official shorthand lol)? my gut feeling is that its better than this.

[tinkady]

oh btw, would a fog shield be a good idea to fit in here to raise the EM rate? in place of a nova maybe?

[~Napalm]

This deck has a PoLdeR index (an appoximate bad draw chance of) 13.4 %. in comparison, SEM has 10 %, standard lava golem rush has 22 %, supernova rush has 17 %. This means that it does not only own, it hardly ever doesnt.

This PoLdeR index is found by approximating good draw odds with a binomial distribution supposing all stochastic variables are independent. I hear ya saying: its hypergeometrical distributed!, yes, but you cannot easily compute the chance of getting a combo of cards with that!! PoLdeR index can.

For SEM i computed the approximate chance of getting only 1 pillar, the real chance is 6.5 %.
For lava golem rush i did the approximate chance of casting a golem from the first 8 cards with the tower + phoenix + immo + golem possibility accounted for.
For supernova rush i did the approximate chance of getting a tower and a supernova first 9 cards, with the automulligan accounted for.
with DEMA it gets trickier, and thats exactly why it rox. I have to add up the possibilities of getting in the first 8 cards one of the following:
E1 immo + ball + golem (60 %)
E2 immo + 2 nova + phoenix + golem (10 %)
E3 2 nova + 2 of phoenix/dagger/gabroid (28 %)
E4 nova + Ball + immo + 2 of phoenix/dagger/gabroid ( 39 %)
E5 2 Ball + 2 immo + 2 of phoenix/dagger/gabroid (14 %)

now dont just add up these percentages!!!!!!!! there is an overlap to the events. Using the principle of inclusion/exclusion (wikipedia it and dont ask) good draw chance is
100*(
(0.60+0.10+0.28+0.39+0.14)
-(0.6*(0.1+0.28+0.39+0.14)+0.1*(0.28+0.39+0.14)+0.28*(0.39+0.14)+0.39*0.14)
+(0.6*(0.1*(0.28+0.39+0.14)+0.28*(0.39+0.14)+0.39*0.14)+0.1*(0.28*(0.39+0.14)+0.39*0.14)+0.28*0.39*0.14)
-0.6*(0.1*(0.28*(0.39+0.14)+0.39*0.14)+0.28*0.39*0.14)
+0.6*0.1*0.28*0.39*0.14)
= 86.555248

  THATS A LOTTA NUMBERS!?!?!?! well it beats answering 100 ppl where i get my numbers from

Win
I'll get back to you on the 200 stats, i'm just over 100 now

[jmdt]

Isn't it nice that we now have choices for EM decks

yeah but we still wanna find the best one. what i'm saying is that while DEMA is pretty great it doesnt EM enough.  It doesnt have enough generation to run more than 3 heals, which is not enough a lot of the time.
how goes testing on christmas (or santa? whats the official shorthand lol)? my gut feeling is that its better than this.

Christmas is on the test list.  There's 4 people testing 4 decks in my group atm, so we'll figure it out eventually.  I also have 1 more EM deck I'll post when I get a chance to.

[tinkady]

sounds good. do you agree with my logic about why christmas should be better tho? or have a flaw to point out?
also ive played a few games with a fog shield and i think its pretty useful, especially against T50. but it might slow it down more than it helps against AI3? and explosions could be more useful vs T50

[tinkady]

Polder, i gave you karma for your very awesome/logical-looking calculations, but are you sure about all your numbers? I play a speedbow a lot, and this seems much less consistent in terms of starting draws than the speedbow does. but I could be wrong, and numbers dont like (as long as theyre the right numbers lol)
also polder have you given any thought to giving different weights to different situations? like based on damage output or just arbitrarily assigned or something
ooh also Polder you said -spark +phoenix increases the index by 4% right? i did that change but i just had a thought...since youre not weighting different situations there maybe a higher chance to get damage out but that damage wont be as high because while theres a higher chance of situations 2-5 in your calculation, theres also a lower chance of getting multiple golems on the first turn or other really good hands.  do you know how this affects things?

[PoLdeR]

I been testing this on my account and Christmas on trainer...and I think christmas is not much slower while EMing almost every time and having more consistent draws. somebody wanna prove me wrong?

yes, getting only an immo has a chance of 6.8 % (exactly) (hypergeometric table + chance of getting 0 tower independent to getting 1 or more immos) which I think is the draw killer for that deck

[PoLdeR]

Polder, i gave you karma for your very awesome/logical-looking calculations, but are you sure about all your numbers? I play a speedbow a lot, and this seems much less consistent in terms of starting draws than the speedbow does. but I could be wrong, and numbers dont like (as long as theyre the right numbers lol)

well thanks for the karma ^^, as for the numbers im sure they are comparable to eachother, are not spot on percentages, more like a polder index, true percentage is about times a factor 5/6 (obtained by some of the truly exact calculations and just hoping this holds for all of them, of which im pretty confident as a mathematician). Typos in math formulas are easy to make, hard to find.... As of now im recalculating the lava golem rush cuz i dont think its right. The more situations you have to calculate the higher deviation i get from the true percentage.
consistency being defined as? 1) you mean getting more bad draws, 2) you mean getting more different games. 1) not true i think, 2) true

also polder have you given any thought to giving different weights to different situations? like based on damage output or just arbitrarily assigned or something

no, thats bound to get messy and unfair, subject to taste and above all you cannot calculate every aspect like take into account the damage total of a situation. I like the idea tho, if you want to do it, just copy paste that big ass number formula and replace 0.68 by 0.68*3 if you think its weight should be 3. What you will get is no longer a percentage at all but a certain number which you can only compare to numbers obtained in the same way. Good luck ^^

ooh also Polder you said -spark +phoenix increases the index by 4% right? i did that change but i just had a thought...since youre not weighting different situations there maybe a higher chance to get damage out but that damage wont be as high because while theres a higher chance of situations 2-5 in your calculation, theres also a lower chance of getting multiple golems on the first turn or other really good hands.  do you know how this affects things?

yea i made a typo in my calculations, its a less than 1 % difference where the slightly less bad draws are in jimmys version. However i think phoenix > Ball, so i play it with the extra phoenix