Guesses Baku/Genn decks even if they haven't activated after mulligan

[justinkb]

Title should be self-explanatory. Ran into this against a Shudderwock shaman. The plugin guessed it was an Even Shaman the first couple of turns, even though Genn definitely hadn't activated. (Guessing Baku decks have the same issue)

[justinkb]

Conversely, when it is definitely a Genn/Baku deck, it also fails to recognize that

[chWolfgang]

Just checking to see if I have missed something? Really hoping this gets fixed. I love this plug in, but it a little less useful when it makes a mistake with odd/even decks.

Thanks for all your effort(s) with this!

[djdookie]

Hey guys, thanks for your comments.
Until now, Advisor solely relies on similarity calculation between HDT's opponent's played (not seen) card list and your archetype deck database to choose the highest match. There is no further logic.

We would need to have some extra logic like:

• Baku was seen (not played) -> drop all decks without Baku for prediction
• Genn was seen (not played) ->drop all decks without Genn for prediction
• Baku was not seen -> drop all decks with Baku for prediction
• Genn was not seen ->drop all decks with Genn for prediction

I don't really like the idea of extra logic for special cards, because Advisor uses a general approach to predict decks. Apart from that this should only be an issue for the very first turns in odd/even decks.

Furthermore I don't have the free time to continue development and I consider Advisor being (core) feature complete (the above is just a convenience feature).

But feel free to improve the code yourself and make a pull request!

[atsepkov]

I came here for the same reason. My familiarity with this project and HDT is only a skin deep from glancing at the source code, but could special-casing Baku and Genn be avoided by treating them the same way as OnOpponentJoustReveal trigger?