AI3 vs AI4, electrum grinding for new players

[kev]

Nice work here.

Might be a good idea to link the Antabow, not only for new players but also because I'm not 100% sure whether you used the build with rares or not.

I think the general feeling is that turns with some decks require further thought and time than others.  I think a lot harder when I'm using a SoSac FG deck than a life rush against AI3, for instance.  Maybe none of the decks listed here require much thought.  If that's the case I was too conservative when I said Essential AI4 matches took 2-3 times as long as Mono-Death AI3 matches, as you show above.

As an aside, Chapuz just mentioned in chat that he'd try to pull together some stats on firecell scavengers so we can figure out where it lies on the table. If I had to guess, I'd say near AI4.

My money's on FGs being substantially better.    Looking forward to Chapuz's conclusions, though.

[ColorlessGreen]

Nice work here.

Might be a good idea to link the Antabow, not only for new players but also because I'm not 100% sure whether you used the build with rares or not.

I think the general feeling is that turns with some decks require further thought and time than others.  I think a lot harder when I'm using a SoSac FG deck than a life rush against AI3, for instance.  Maybe none of the decks listed here require much thought.  If that's the case I was too conservative when I said Essential AI4 matches took 2-3 times as long as Mono-Death AI3 matches, as you show above.

As an aside, Chapuz just mentioned in chat that he'd try to pull together some stats on firecell scavengers so we can figure out where it lies on the table. If I had to guess, I'd say near AI4.

My money's on FGs being substantially better.    Looking forward to Chapuz's conclusions, though.

I used the two-arsenic version of antabow, and it's linked in the arena stats thread already, but just for convenience:

Spoiler for Hidden:

Hover over cards for details, click for permalink

Deck import code : [Select]

4sa 4sa 4sa 4sa 4vh 4vh 4vj 4vj 4vj 4vj 4vj 4vj 52q 52q 55q 55q 590 590 590 590 590 5c1 5c1 5f6 5f6 5i7 5og 5og 61q 61q 8pm

About firecell, IIRC liquid antimatter had a FGEI of around 2k, and my general feel for firecell is that it's less profitable (mostly due to being slower and having less skips). I'm really looking forward to actually finding out what firecell's FGEI is, though.

About time-per-turn: I'm willing to say that monodeath, monoaether, and antabow are all close enough to the same amount of time to play per turn, though antabow is slightly slower due to having to consider whether/where to use CC/PC. I don't think that (after enough games to be comfortable with the deck) it's really a large enough difference to matter in this context, though.

In a larger context, there's definitely some stallier decks or decks with randomness/lots of draw power/lots of creature spam/etc that slow down the turns as compared to the play-what-you-draw style here. The formula I'm using doesn't directly account for that, but it's pretty easy to come up with some assumptions when comparing two numbers along the lines of "hmm, i'd say it takes me 25% longer to play pdials than monoair, so i'll multiply the monoair rate for 250 turns by 125%". I find that all to be greatly varying between different players, though, so I don't really have any ideas on incorporating it in to the formula beyond a subjectively-chosen multiplier at the end.

edit: LA had a FGEI of 2281, so there's plenty of room for firecell to be less profitable than LA and still more profitable than AI4. Looking forward to finding out.

[Marvaddin]

Ok, I have finished my 1st 200 games using TADAbow vs AI3.

Spoiler for TADAbow deck:

Hover over cards for details, click for permalink

Deck import code : [Select]

6qq 6qq 6qq 6qq 6qq 6u3 6u3 6u3 6u3 6u3 6u3 6u5 6u5 6u5 713 74f 74f 74f 77g 77g 7ai 7ai 7ai 7ai 7dq 7gm 7gm 7jr 7tf 80g 8pj

Results:
200 Games
194 Wins
6 Losses (-60 electrum)
154 EMs (154 x 40 = 6160 electrum)

Non EM victories rewards were:
19 - 16x = 304
18 - 9x = 162
17 - 5x = 85
16 - 4x = 64
15 - 1x = 15
14 - 2x = 28
13 - 3x = 39
Total: 697

Note: In your calculations, Kev, you did number of games x10, but this is the minimum prize (in fact, I think its 11). If you divide 697/40 games, the media is 17,4 (i know, my deck is upgraded). Even if you consider it equally for both AI3 and AI4, it seems inaccurate, because winning with low HP is much more frequent in AI4.

Cards sold: 4329 electrum

Spoiler for Cards Sold:

1 Aether Pillar: 4
1 Dim Shield: 42
1 Phase Dragon: 77
1 Phase Spider: 19
1 PU: 43
1 Lobotomizer: 147
5 Bonds: 210
6 Firefly Queen: 420
4 Owl Eye: 596
2 Hope: 86
1 Shockwave: 38
4 Voodoo: 152
1 Nightfall: 39
1 Pandemonium: 39
1 Vampiric Stiletto: 145
1 Drain Life: 39
1 Soul Catcher: 16
1 Flesh Spider: 19
1 Mutation: 38
1 Plague: 40
1 Stone Skin: 38
1 EQ: 39
2 Basilisk Blood: 76
4 Deflagration: 76
1 Rage Potion: 39
1 Dusk Mantle: 42
1 Phoenix: 71
1 Gravity pull: 18
1 Momentum: 18
2 Gravity Shield: 82
1 Otyugh: 40
1 Armagio: 41
1 Cockatrice: 19
1 Fallen Elf: 68
1 Max Demon: 41
1 Horned Frog: 18
2 Sky Blitz: 90
3 Blessing: 117
1 Guardian Angel: 19
2 Pegasus: 80
1 Wyrm: 20
1 Crusader: 41
2 Scarab: 36
1 Pharaoh: 265
1 Hourglass: 40
1 Heal: 19
1 Druidic Staff: 146
1 Toadfish: 21
1 Morning Star: 149
1 Reverse Time: 18
3 Steam Machine: 120
3 Rain of Fire: 213

Total: 82 cards, 4329 electrum

Double spins: 98x (no card) + 82 (cards) = 180 x 5 = 900

Total: -60 + 6160 + 697 + 4329 + 900 = 12026 electrum
12026 / 200 = 60,13 electrum / game.

Almost the double of your calculations for unupgraded, but considering that you ignored double spins, and your reward for no EMs was most probably underestimated. Pretty good also considering that most games last for like 1 minute or less.

Well, this is what I have for now, maybe I will extend the tests more later 8)

[Tiko]

I was wondering how the simulations work out in practice, but the fact how slow and annoying Essential can be with the improper draws made me try from a different view. So I thought I'd record a small (50games) sample of my regular Half-Blood Hunting Hour, just to be sure the payout is worthy enough for anyone to bother.

I used the following deck:

Spoiler for PvP Tsunami:

Hover over cards for details, click for permalink

Deck import code : [Select]

6rk 6rk 7gk 7gk 7gk 7gk 7gk 7gk 7gm 7gm 7gm 7gm 7gm 7gm 7gp 7gp 7gq 7gq 7gt 7gt 7gu 7gu 7gu 7gu 7i6 7i6 7i6 7i6 7i6 7i6 8pp

As you see, it's a simple mod of the classic mono-Water rush that I use from time to time to harass random PoisonDialers in PvP2. As it is fully upgraded, it's much faster than Essential, but in exchange, its limited defense is not suited for long-term stalling. Makes it even enough I guess, to post all this here..

And I ended up with the following stats:

Spoiler for Stats:

W/L-TTW-Electrum-Up/Rare

W 10  29
W 13  28
W 8    31
W 12  34
L
W 11  32 Seraph+
W 11  21
W 8    19
w 10   30 Gnome Gemfinder
L
W 9    31
L
W 10  36
W 8    23
W 8    38
W 15  33
W 8    42
L
W 7    29
W 14  15
W 12  29
L
W 10  35
W 9    26
L
W 8    33 Innundation
L
L
W 10  36
W 9    29
W 10  28
W 8    85*
W 10  39
W 8    29 Nymph's Tear+
W 12  25 Spine Carapace
W 8    41
W 8    42
W 10  44 Amethyst Tower
L
W 12  20
W 9    39
W 9    35
L
L
W 9    41 Arsenic
W 13  17
W 9    27
W 11  32
W 9    32 Eternity+
L

W%=76% (38/50), Avg. TTW: ~10turns (9,8), *EMs: 1, Electrum gain: 995(+8134), including matching spins and ups, excluding any common card won.
I know that somewhat matching results can be achieved with any other elements' similar build. It's personal preference.

Conclusion: So, in about 2x2hours throughout the weekend, playing sporadically mostly in coffeebreaks (in other words: without giving much effort); after selling all the excess cards, about 10k electrum and another white pixel's worth of score has been added to my account. Not bad from a simple mono deck, and the 50hp drop on HBs is just 'icing on the cake', and makes it all-round easier and newbie friendly for everyone.

I still say HBs do worth the effort, and not only their payout turns out to be pretty nice compared to the common stereotype about them, the game remains far more entertaining on the long run than any Elder/Arena-grinding session, thanks to their awesome variety. Though the spins can be quite insolent, and of course you have to sell everything you earn to make them fully profitable, HBs are always there to remind you that the game is about fun and who knows? You may always learn a new trick from them :]

[TribalTrouble]

Well said Tiko.

[Marvaddin]

I also tried a 50 games run vs AI4 this Sunday.

This was my deck:

Spoiler for High Gravity Bronze:

Hover over cards for details, click for permalink

Deck import code : [Select]

568 744 744 744 744 744 744 745 745 745 745 745 745 749 74b 74b 74b 74b 74c 74d 74d 752 752 752 75m 75m 75m 75m 75m 75m 8pl

Results:
50 Games
46 Wins
4 Losses
21 EMs

I guess the luck was on my side I wouldnt expect 21 EMs (although sometimes I delayed the game end for 1-2 turns to try an EM). As its a control deck, 12,1 TTW.
Electrum rewards (including double spins): 2628, -80 (losses) = 2548, which is 50,96 / game.

Cards obtained is highly variable, but I was also lucky. I got 5 upgraded cards (Firefly, Ice Lance, Improved Fog, Chimera and Improved Steal), and 5 basic cards (Bonds, Quint, Flooding, Horned Frog and Trident). Surprising, because basic cards drop rate is expected to be higher. 6081  Electrum from cards.

Total: 8629/50 = 172,58 electrum / game.

Did you get 7 upgraded cards in 38 wins, Tiko? Really? Wow, Kev have tested 300 wins, and the drop rate was about 7,7%. Guy, you were lucky this weekend

Hmmm, anyway, with a high win rate (92%), high EM rate (42%), and a good upgraded card drop rate (10,9% per victory), it was less than 3x what I got vs AI3. And playing is more than 3x longer. Not only like double of turns, but you also need to think about targets for SoFo and Oty skills. So, I wouldnt really be sure about AI4 farming being more profitable than AI3 farming (which is also much less luck dependant).

Of course I agree that playing AI4 is much more fun. Before my 200 games vs AI3, I wasnt playing AI3 for probably several months. And I also like playing AI4, I play them a bit almost everyday. What its unknown, and we will probably disagree, is which one is best for farming electrum.

Probably, farming FGs is the best, but to make it efficient (even more using unupgraded decks, or 6 upgrades decks), you need skip a lot of fights. And I hate to do that, because I care about score and win rate. About AI3 vs AI4, I would probably go for AI3. As my sample shows, its possible to get 60 electrum per game once you have an upgraded, efficient EM grinder deck (with games lasting about 1 minute). Because of this, I recommend to upgrade AI3 farmer before FG farmer. After that, if you need you can have like 3000+ electrum per hour of grinding, 2 upgraded cards for any deck you want to create.

Conclusion: I wouldnt go for FGs unupgraded (of course, people are free to do this). To farm electrum, I would do AI3. To have fun, surely not AI3. AI4 is ok if you want a mix of electrum and fun (although I think its hard to have fun using that mono aether deck).

Edit: upgraded cards drop rate for AI4 can really suck. I wasnt taking notes, but after +40 games vs AI4, with a 36-4 record, just +1 upgraded card (Precognition).

[Tiko]

Did you get 7 upgraded cards in 38 wins, Tiko? Really? Wow, Kev have tested 300 wins, and the drop rate was about 7,7%. Guy, you were lucky this weekend

Edit: upgraded cards drop rate for AI4 can really suck. I wasnt taking notes, but after +40 games vs AI4, with a 36-4 record, just +1 upgraded card (Precognition).

Lucky indeed, it was an exceptionally generous droprate. But I'd say (after almost 10k games against them) over the course of ~50games the overall average of ~5 upped cards is almost assured. Sometimes more, sometimes less. Just ignore the similar un/upped spins, and enjoy.