I thought of creating an entirely new topic, but instead decided to post it here because it's not like it needed discussion or something, and this being the Halfblood Stats Thread made it appropriate to be here.
I decided to test the TTF versus AI4s to understand how fast a deck, approximatively, needs to be if it has no protections in order to overcome AI4s.
I used XenoSim with the following deck:
Spoiler for deck:
Hover over cards for details, click for permalink
6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 6qq 8pl(I'll explain later why I used this decklist)
I ran
19973 games (more on the number later) and it resulted into an average of
9.579282030741501 TTF. Only for the record, 19115 of those games were EMs, meaning that 858 weren't EMs.
I used that deck because:
1.It needed to be inoffensive, of course;
2.Needed to feed Devourers if AI played them, I decided that it would have been a better choice than to not feed them at all, because it may have altered the speed of the deck;
3.Shouldn't, as much as possible, lose due to deckout, because the nature of the study is to how fast a rush deck needs to be in order to kill AI4s, if you lose by deckout the whole point in that would be missing, and opted to resist to deckouts.
The number of game may look strange due to two things:
1.Low number of games (lesser issue), but I ran the game 1000 at time and it looked like the data was constant, so I decided to stop there;
2.Not a rounded up number of games (major issue).
It's not like I ran 19973 games, I already said I got the data 1000 at time, so it should be 20000 if anything. Indeed, I DID run 20000 games, but the AI lost, you read that right, LOST, 27 games. I decided not to round that up by adding other games.
Full stats:
PlayerWins: 27
Winrate: 0.135 %
Average TTW: 11.851851851851851
EMs: 3
EM rate: 11.11111111111111 %
ComputerWins: 19973
Winrate: 99.865 %
Average TTW: 9.579282030741501
EMs: 19115
EM rate: 95.70420067090572 %
I thought XenoSim had some issues while reading those stats, but I managed to explain something of that:
The 838 non-EM AI's wins couldn't be explained with SoSac, given that, as far as I know, Shards aren't allowed into randomization of the decks. I then gave it some thought and realized that, mostly, those non-EMs may be explained with Singularities which can, and most likely will, get vampire on top of their antimattered attack. A little part of that may be due to PU'd dolls on my side of the field because of Pandemoniums. That's unlikely, though.
As for my 27 wins, they must have been due to deck out. I couldn't come up with another excuse (aside a PU'd Doll with an high amount of poison on it or something like that, expecially given the relatively low TTW) if not a deckout due to Time decks overdrawing without having Eternity or enough quanta to keep an Eternity loop. I find Singularities kills to be highly improbable due to the AI having to play either two SNs or three Novae and not winning after that much. However the EMs may be explained by Chrcord.
Aside those strange numbers, I think I can finally say that a deck which can perform 8-turns kills and has some PC, possibly, can do well in AI4.
