Why does it choose to:

834. attackWithMinion: Leper Gnome target: Ooze choice: 0 place0
943. attackWithHero: Malfurion Stormrage target: Knight of the Wild choice: 0 place0

And not the other way around? Now it takes 6 facedamage from the Knight of the Wild, instead of 1 from the Ooze.

Data sync. Look at the action list after the first board of the turn to see what SF really wanted to do.

Druid hero power is totally screwed because of data sync. HR usually tells SF that your hero doesn't have any attack right after shapeshift is used, then on the next update it does.

So I should be using a standard capable hunter or shaman deck to hit 5? I was using a Wild Secret Paladin deck because I remember reading on here that it's computationally less intensive than other decks which means the bot will rope less which is obviously crucial. But if there's 4x as many wild decks as standard, I can see why it's not getting above rank 10. The issue there is that a new set is being introduced tonight and I don't have all the cards for hunter, only shaman lol. Thanks for the help

If you modify the AI to use speed it may still encounter datasync when HR sends bad data. But yes, datasync is a problem with HR collecting data about the board too early.

As I've said many times, SF does not care about finding better lethals. Lethal is lethal.

To make sure that any mis-play is a real mis-play you must first change these two variables in settings.txt to false:

playarround = true;
useSecretsPlayArround = true;

If it does it again with these set to false, then it may then be a SF issue. I hope this helps :)

@sepefeets I can't seem to find a way to force Raven Idol (LOE_115) to select Discover Spell, any chance you can help?

Hmm I was hoping you weren't going to say that, I have tried both, even to the extreme of 500. I'll double check everything again and get back to you :)

Both logs are the same and do not have any of the plays you mention. They all sound like they are leaning on the micro-optimization side that I warned you about though.

@sepe (or anyone who knows)

Question for the new mulligan. could someone elaborate the differences between those 2:

hold;shaman;all;OG_024:1:CS2_106,0,0+CS2_222,0,2
\\hold 1 copy of OG_024 if you are shaman vs any class and have 1x CS2_106 and 2x CS2_222 but don't hold any CS2_106 or CS2_222 (other rules may hold them before/after this rule)

hold;shaman;all;OG_024:2:CS2_106,2,2+CS2_222
\\hold up to 2 copies of OG_024 if you are shaman vs any class and have 2x CS2_106 and 1x CS2_222 and hold the 2x CS2_106 and 1x CS2_222

Or other said:

Does the first rule means it discards any CS2_106 and CS2_222 even if it has OG_024?

Or other said:

Does the 2nd rule means it holds 2x CS2_106 and 1x CS2_222 even if it has no OG_024?


Sorry for the noobisch way to ask this, I'am pretty sure i use it correctly but I dont fully understand the possibilites of these, and i want to know to improve the mulligans to the fullest.

So first of all, holding/discarding is dynamic. By default everything is discarded if you have any custom rules, then the 1st rule is processed, then the 2nd, 3rd, etc. The 1st rule can say hold 2, then the 2nd rule can say hold 1 or hold 0, and 2 will still be held. However, if rule 3 says discard 2 then it will discard 2 even though the previous rules wanted to hold some. If the final decision is to hold 1 then it implies to discard the next if there are 2 copies and vice versa with discard 1 implying to hold the 2nd.

So with that said, the 1st rule quoted will discard CS2_106 and CS2_222 if no other rules have changed the default action for them. What "hold/discard 0" really means is "don't change whatever action was previously set."

For the 2nd quote rule, requirements are only processed if the main card exists in the mulligan so it is saying you must have 2x CS2_106 and 1x CS2_222 and OG_024 to hold them all.