Elements the Game Forum - Free Online Fantasy Card Game
Elements the Game => General Discussion => Farming Studies and Statistics => Topic started by: Ungaros on December 09, 2013, 05:49:22 am
-
I ran each of the FG decks 50k times in Xenocidius's Slot Machine Simulator (http://xenocidius.webs.com/ (http://xenocidius.webs.com/)) and recorded the cards/win:
Divine Glory 0.675
Decay 0.616
Jezebel 0.598
Octane 0.598
Eternal Phoenix 0.554
Hecate 0.548
Akebono 0.526
Osiris 0.514
Fire Queen 0.511
Gemini 0.498
Incarnate 0.486
Serket 0.485
Obliterator 0.484
Paradox 0.480
Lionheart 0.477
Graviton 0.470
Neptune 0.470
Destiny 0.450
Ferox 0.445
Seism 0.415
Dark Matter 0.409
Chaos Lord 0.391
Hermes 0.391
Miracle 0.377
Dream Catcher 0.376
Elidnis 0.374
Morte 0.374
Scorpio 0.366
Rainbow 0.360
I used the deck codes found here: http://elementscommunity.org/forum/false-gods/here-are-the-decks-of-the-actual-false-gods/ (http://elementscommunity.org/forum/false-gods/here-are-the-decks-of-the-actual-false-gods/)
-
Thanks a bunch, now I don't have to go digging for them every time this comes up in conversation. +rep
edit: average of shown numbers is 0.473
-
It seems these stats should be used to increase the accuracy of FGei and your Uei stuff, right? Decks that defeat FG's towards the top of this list and have skips tending toward the bottom of this list should get boosts and vice versa.
-
It seems these stats should be used to increase the accuracy of FGei and your Uei stuff, right? Decks that defeat FG's towards the top of this list and have skips tending toward the bottom of this list should get boosts and vice versa.
These stats essentially match darkweaver's stats, which were used in most of the versions of the original FGEI studies and the statmasta spreadsheet. I don't recall if Pella got around to updating the current FGEI study with these numbers instead of the all-god average, but whatever FGEI study gets run for 1.4 will include updated drop rates.
UEI isn't generally used for FG so this isn't really useful there. Also, UEI doesn't have a mechanism for normalizing or for using this information, so FGEI should continue to be used for FGs IMO. UEI already uses results from xenosim for the drop rates for ai3 (and assumes the regular spins from arena approximately match the spins from ai3).
-
Does the simulator contain the SoFo cards added to Akebono and Osiris ? The added cards slightly alter the winrate for those 2 FGs, so I just wonder if the above table is up-to-date for that change.
-
The simulator just takes a deck code and spits out spin results. Ungaros was using the thread containing the current deck codes, so unless he typoed something, these should be accurate for patch 1.32.
-
Does the simulator contain the SoFo cards added to Akebono and Osiris ? The added cards slightly alter the winrate for those 2 FGs, so I just wonder if the above table is up-to-date for that change.
Yes, those updates were indeed included, and I did observe an increase in the drop rates of Akebono and Osiris when compared to DarkWeaver's stats.
-
I find it odd that paradox is in the middle when mathematically it should have the highest yield. The rest look fairly on point though.
-
Even if Paradox has a lot of copies of some cards, it also has a lot of non-quanta cards (towers are nearly negligible because they are heavily disfavored in the spins).
Paradox: 8 RoL, 8 Blessing, 8 Deja Vu, 5 Twin Universe, 4 Hourglass, 3 Miracle, 2 Mirror Shield, 2 Morning Glory (Total: 40)
Divine Glory and Decay are at the top because they have lots of copies of some cards, but also because they don't have very many non-quanta cards.
Divine Glory: 6 Explosion, 6 Glory, 5 AW, 3 Miracle (Total: 20)
Decay: 6 Pest, 6 Siphon Life, 6 Fractal, 2 Steal, 1 Dusk, 1 Eclipse, 1 Electrocutor (Total: 23)
We can also compare some other FGs that rate lower but still above Paradox:
Akebono: 6 Massive Dragon, 4 Overdrive, 2 Armagio, 2 Chimera, 2 SoFo, 1 Tower Shield, 1 Unstoppable, 1 Titan (Total: 19)
Gemini: 6 TU, 6 Phase Recluse, 4 Unstoppable, 4 Phase Shield, 3 Massive Dragon, 1 Immortal, 1 Lobo, 1 Phase Dragon (Total: 26)
Even if Paradox has more copies of cards, those copies are still diluted because of the large number of non-quanta cards.
-
I thought it still took a sample of the deck though...
I've also almost always have gotten a card from paradox...
Hmmm... well, I guess that's why we test, just seemed odd to me.
-
It chooses several cards at random from the deck and then spins among those. The chance of getting a card will be higher if some cards are chosen multiple times, and a card that makes up a large proportion of the deck is more likely to be chosen multiple times.
-
These stats essentially match darkweaver's stats, which were used in most of the versions of the original FGEI studies and the statmasta spreadsheet.
Yes, very much so.
Only Akebono and Osiris suffered a notable drop (>5%) because they each got two SoFs added to their deck after DarkWeaver calculated the stats.
blarg: