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

[CrockettRocket]

Keep in mind we don't test FG killers like that anymore (Random X number of games). Instead, you're suppose to do 10 games against every FG. Then to say which ones you should skip is personal taste on your reccomendation, I believe most people skip if the winrate is less then 30%.

[Lastmerlin]

Well for a very large number of games these two methods should yield the same result. Of course, my method has more variance. Source of variance are next to your draws and the FG draws (which already are a huge gamble, as mentioned this decks mainly looses to bad draws) the random FG you draw. The method you propose eliminates this factor and therefore yields better results with less games. But at the end, for very large numbers both methods converge to the expected winrate. Moreover, the main random factor with this deck is rather your own draws (there is a large number of FG that just loose equally if you get your combos nicely), which is not excluded.

Hence, this data is nowhere completely worthless. I believe, that lets say 2500 games from 10 players provide a far better result than one sweep test as you proposed, because it also averages out factors like player experience and playstyle. Now the final trick is: Player play this sort of decks for lots of games anyway because the game is so grindy. All you need is to collect the data. Its an additional 5 seconds per game to make a list. The direct selection of the FG requires the use of a trainer, hence the games earn you nothing. This is at least 100 seconds *lost* per game. In other words: The sweep of 290 games in the trainer is more expensive in terms of additional time than the 2500 random games. If anyone is motivated to do that, go ahead. I will rather produce data as a side effect of gaining money. If a few other players do the same the results should be very useful.

[Narukta]

Well, you could just keep track of the gods you've played against, calculate winrate for each of them and then average.

For example: I've seen Scorpio 15 times and I won 12 games, Lionheart 3 times ad I won 0 times, Jezebel 1 time and I won 1 time, Ferox 10 times and I won 3. The other gods are dead. So the winrate for each god is 12/15, 0/3, 1/1, 3/10 and you expect to meet each god in 1/4 of the all matches, so the probability to win is 1/4*(12/15+0/3+1/1+3/10)= 21/40, so the total winrate is 52,5%.

In this way you can both farm on your account and have more precise data.

[nolf]

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.

Cheers for the test. Slightly disappointing. Your observations are definitely true. I sometimes win by unorthodox methods, using one or two or even all of those precious TUs to copy FG creatures... for me, variation is a plus, rather than going through the same motions every time (which becomes stale after a few thousand matches). When I feel like playing some more, I visit the arena with a mutation or shard of serendipity deck. Not very efficient, but different creatures to play with every time.

[Keeps]

I always thought it was understood this deck wasn't a >50% WR deck, it was just about how fast it played that made it the highest profit Per Hour Deck is why people played this.

[Manuel]

won against dark matter with a singularity from turn 3 in the field (i was thinking that i was using the nova mod)

Spoiler for Hidden:

give me an award  8)

[eaglgenes101]

won against dark matter with a singularity from turn 3 in the field (i was thinking that i was using the nova mod)

Spoiler for Hidden:

give me an award  8)

Here, choose a cookie:
https://encrypted.google.com/search?q=cookie&source=lnms&tbm=isch