Voodoo bravery [FG][Arena][Fast][Focused]

[Lastmerlin]

After familiarizing with the deck, I did some statistics with the deck variant proposed in the False Gods with Stats thread. What I got with 250 games was:
140 wins (56%), 110 losses (44%)

Apparently, I am clearly below the claimed 64% winrate that is claimed by several posters here. The false gods thread by dictator lists 58% winrate for the small skiplist (6 gods) which is much closer to my result. Any ideas where these differences come from?

I also tracked the specific reasons of loss and got:
no pendulum: 7 (6.4%)
no nova: 20 (18.2%)
no doll: 14 (12.7%)
no gravitiy pull: 21 (19%)
insufficient basilisk blood: 10 (9.1%)
insufficient TU: 7 (6.4%)
Other: 31 (28.2%)

Other contains all other reasons like too much damage, momentum, poison, damage burst kills your doll or whatever. The two categories BB/TU are a bit unsharp. because often a BB or a TU can do the job. These categories only count losses where the combo essentially fails due to the severe misdistribution of one card type in the deck. Any situation, where you can play the combo, but you are just too slow or lack a specific additional card (like the last BB) is noted in Other. What you can see is that the deck mainly really looses by shooting into its own foot. The number of possible RNG trolls is astonishing. The biggest culprits are nova and gravity pull. Therefore I added one of each to the deck:

Hover over cards for details, click for permalink

Deck import code : [Select]

6u3 6u3 6u3 6u3 6u3 6u3 6ve 6ve 6ve 6ve 6ve 74d 74d 74d 77l 77l 77l 77l 77l 77l 7ee 7ee 7ee 7ee 7ee 7tg 7tg 7tg 80b 80b 80b 80b 8pu

Yes, that not exactly 30 cards anymore. I dont believe that exactly 30 cards is a must have - fine tuning the relative frequency of the cards might outweight the benefits of minimal deck size. If I must take something out, I would choose a SoB and a BB which leads to exactly the deck of Submaschine. But as I said, I am not really convinced about that.

I did a second run of 250 games and got:
144 wins (57.6), 106 losses (42.4%)
With following reasons for the losses:
no pendulum: 10 (9.4%)
no nova: 13 (12.3%)
no doll: 12 (11.3%)
no gravitiy pull: 21 (5.7%)
insufficient basilisk blood: 11 (10.4%)
insufficient TU: 10 (9.4%)
Other: 44 (41.5%)

Overall, I hoped for a more significant improvement. I was above 60% winrate almost all the time until an infuriating streak of rng troll losses during the last 50 games ruined it. In my opinion the additional gravity pull is a big improvement. It reduces this factor of loss greatly and allows much more versatile strategies, like copying a doll one turn earlier if you suspect a damage bust and pull the copy for the last bit of damage. Apart from that, this variation plays more pleasant, because you get the combo out more often. I am not completely sure about the additional nova. Of course, it apparently helped to reduce this reason of loss, but its still the leading cause (apart from Other) and dilutes all other cards while increasing the number of novas by only 20%.

Under the viewpoint of raw efficiency, it might be worse, because the reasons of loss are shifted towards the later components of the combo. That means that your lost games take more time. For maximum efficiency I believe that could even make a 5 pend, 5 nova version and skip any starting hand without pend and nova. My experience says, that games without pend, nova in the starting hand are mostly  lost anyway. The chances to draw one in time are often overestimated and then you are sitting there for turns drawing and discarding...

[Manuel]

@Lastmerlin

i can't win a single battle (after 10 battles on the trainer)

is not so simple draw at least one  pendulum in the first hand (or even after two draw); i don't know if u was really lucky to lost only the 6,4% of the match for "no pendulum" or if u posted the wrong version, or i am very unlucky

edit: win 2 in a row, and another lose against a god beat five minutes ago because i didn't draw a pendulum after 2 draw (and only one shard after six turns, maybe i am unlucky)

[omegareaper7]

@Lastmerlin

i can't win a single battle (after 10 battles on the trainer)

is not so simple draw at least one  pendulum in the first hand (or even after two draw); i don't know if u was really lucky to lost only the 6,4% of the match for "no pendulum" or if u posted the wrong version, or i am very unlucky

edit: win 2 in a row, and another lose against a god beat five minutes ago because i didn't draw a pendulum after 2 draw (and only one shard after six turns, maybe i am unlucky)

With 5 pendulums in a 32 card deck, your chances of not drawing any are extremely slim. It was just bad luck on your end.

[Higurashi]

@Lastmerlin: It may be because of the second factor when it comes to farmin'. The first is to know your deck and the second to know your opponent. In "Other" you're listing burst damage, which is something you can only get an intuitive sense for after hundreds of battles against the same FG. This, more often than not, denotes a big winrate difference between users. For some other decks like ye olde Eternal Concordance you had winrates differ with as much as 30%.

If not that, it may just be statistical variance. You're going to need some 10k games before stat percentages start to become somewhat reliable.

[Lastmerlin]

@ Higurashi: Of course you are right - more experience would have won more games. But I estimate the margin to be only about 3% (which puts me on par with the 58% given in the FG farming stats thread). The idea of the "Other" category was to separate bad luck on your side (which you can try to optimize by improved card ratios) from good luck on FG side (which you cant influence). The Other category includes many cases, where you can't do anything either, like: Decay never playing Eclipse, Octane playing only Gas, Scorpio/Morte playing only poison, Gemini/Graviton putting out 5+ momentum creatures, and just massive damage that nothing you could do can handle (e.g. Hecate). The cases where you can actually decide something are rather seldom. I would estimate one third of the *Other* cases, which is 10%. Within these 10%, an improvement potential of 30% seems realistic.

An interesting remark: I consider my version having more improvement potential for experiences players, because share of "Other" is increased and there are more cases where you do something interesting. With three GF you can do things like copy your buffed doll early, GF a plain doll for one turn extra delay and then return to GF your copied doll, hopefully with one additional BB to somehow survive and get the damage in.

After all, this is exactly the reason why I post my stats and ideas here: Because I hope that some more experienced players give these modifications a try and report whether they are helpful.

@omegareaper: If my calculations are correct, the chance to have no pend in the first 7 cards is 26.38%. Due to automulligan, you need the bad luck twice, which yields 6.96%.  My second batch is above this number because I just had bad luck - as I reported I had a real streak of such events. For a lot of games it should be even below this, because you can survive if you get the pend with your first three draws. For this tests, I played until I really lost - but for efficient play I suggest to give up if you start discarding non-redudant cards or your last SoBra.

[eaglgenes101]

The main variants of Voodoo Bravery are designed to draw quickly, and run through the deck reliably. Hence slim building and lots of Shard of Bravery. As testing has revealed, even small changes in the deck can have an unexpectedly significant on the chances of getting the cards you need, so I wouldn't be surprised if your build's weight is holding it back. But I won't be too critical. Empirical evidence > theory, and we're still collecting the former.

[Lastmerlin]

Well, before critizising and speculating why my proposal might be bad - please note that I did provide empirical evidence. And my results say, that its as least as good as the original variant. The batch of 500 games overall is not huge, but quite considerable for one person. Nevertheless, the difference might be too small to be statistically significant (perhaps I make the calculations some day, its not very difficult). Thats why I posted it here, because I hope that other will run some tests (ideally players more experienced with the deck) and report some statistics.

Concerning the overall win rate: The big difference is between the FG farming thread (which reports 58%) and some early posts here (64%). I would really like to know if anybody can reproduce the 64% nowadays. I somewhat suspect, that changes with the last version of the game (AI modifications or whatever) changed winrate because the high win rate stats are always reported before the date 1.32 and the newer stats are always lower. This is a test completely independent of my proposed slight modification.

[CuCN]

The only post in this thread that claims 64% is this one, which doesn't include skips.

[nolf]

@Lastmerlin: Nice to see you've been doing more testing and adding another variant.

I still log in most days, just spin the Oracle and play three quick games vs FGs with the (mostly useless) creature the Oracle provided. You have to hand it to the deck, the win rate is good and it's just damn fast to play, over within a few turn. Having played thousands of games against FGs with Voodoo Bravery I think it's true experience helps, allowing you to win in unconventional ways at times.

Personally, I settled on this variant:

Hover over cards for details, click for permalink

Deck import code : [Select]

4vj 4vj 4vj 6u3 6u3 6u3 6ve 6ve 6ve 6ve 6ve 74d 74d 74d 77l 77l 77l 77l 77l 77l 7ee 7ee 7ee 7ee 7tg 7tg 7tg 80b 80b 80b 8pu

[Lastmerlin]

OK, then just lets assume, that 58% is a more realistic aim for an experienced player and stop worrying about this a bit dubious higher values.
Then testing variants is the next step again.

So thanks to nolf - perhaps I will use this one for the next batch (If I get the time). Essentially you did, what I did not dare -taking something out for the additional GP and nova.  If I see it correctly its one TU and von SoB. The really interesting thing is the nova split. Apparently this can help to get a faster start. The question is: How often does it do the trick. For me, getting the doll and the quanta to play it is the essentiall step, because this is point where you actually can start quickdrawing and use these shards for good. I will give it try.

[nolf]

Give it a try and see what it feels like. I made these changes over time, like you I felt I should definitely have six SNs, three dolls and three GPs. And then I came up with that nova split and am quite pleased with myself, it does a pretty decent job of combining the advantages of both OP versions. I believe it is more reliable, that is, there's more chance of getting the "combo" started. But that might not translate into actually doing better in a test - maybe I get the combo started more regularly, but don't win more often. At least it is more satisfying to actually be doing something, rather than having a clogged up hand.

[Lastmerlin]

@Lastmerlin: Nice to see you've been doing more testing and adding another variant.

I still log in most days, just spin the Oracle and play three quick games vs FGs with the (mostly useless) creature the Oracle provided. You have to hand it to the deck, the win rate is good and it's just damn fast to play, over within a few turn. Having played thousands of games against FGs with Voodoo Bravery I think it's true experience helps, allowing you to win in unconventional ways at times.

Personally, I settled on this variant:

Hover over cards for details, click for permalink

Deck import code : [Select]

4vj 4vj 4vj 6u3 6u3 6u3 6ve 6ve 6ve 6ve 6ve 74d 74d 74d 77l 77l 77l 77l 77l 77l 7ee 7ee 7ee 7ee 7tg 7tg 7tg 80b 80b 80b 8pu

I tried this variant by nolf for another 250 games. After this quite thorough testing, my results say: This is definitely no improvement.

I struggled all time to stay above 50% winrate and ultimately failed with final result 123 wins, 127 losses.

Specific reasons of loss:
no pendulum: 6 (4.7%)
insufficient nova: 28 (22.0%)
no doll: 12 (8.9%)
no gravity pull: 13 (10.2%)
insufficient basilisk blood: 7 (5.5%)
insufficient TU: 6 (4.7%)
Other: 55 (43.3%)

Observations: These unupped novas did really no good for me. The main issue is that they mess up the mulligan. Your starting hand really often has nova + supernova oder nova + pendulum and both often turn into a mess. The nova gives you a chance to draw once per SoB for another nova or SN, but this yields nothing surprisingly often. In the stats this is reflected by the increase in the nova category (which includes now just one nova and nothing else, because this easily kills you as well).

The deck is much sharper that the original. I often won or lost by a few HP on the FG or the doll. Here, you often have to gamble somewhere (like, will the FG put out at least X or at most Y damage?) and the lost gambles ramp up the loss rate. Issues are two things: First, the reduced quanta supply by the novas is really noticable. I often lost because I could not afford to play the fourth BB from 1 Nova+ 1 SN. Seconds, the fourth TU is really missing. Having less TU  than you would like to have is one of the main reasons for these gambles.

The decks succeeds at providing more often a chance to do *something*.  These novas often provide fuel to play around a bit, delay with a single BB or whatever. Hence it is more difficult to put lost games into one of the categories clearly. I should have marked much more games as lost due to TU, because lack of these often forces you to gamble.

Overall, this decks most probably more potential for improvement when played by someone very experienced with this specific variant because you learn which risks to take. However, I doubt it can get close to the winrates of the other versions.