How much Depth does KeyForge have?
A while ago I wrote an article that touched on the Depth in KeyForge. In response, George Keagle got excited and asked me to explore further. The problem was I needed a good example, and some statistics, and I just didn’t have those. I do now. And I am very excited to share it with you.
A new player was recently pointed at my article for the top 5 heuristics in KeyForge and hit some problems. Specifically, they felt that following the Hand + Board heuristic was limiting. After some discussion, people proposed that maybe new sets don’t hold up to the same heuristic. I took the opportunity and went to review how many turns I went with hand + board in my recordings from the Swindle “The Big One 2” tournament, which was DT only Archon solo. See, I didn’t want to go into a game with the intention of counting hand + board because that might bias my decisions.
I counted a total of 45 turns.
29 or 64.4% of which I went hand + board
8 or 17.8% of which there was an even choice and I went with one of them.
8 or 17.8% of which I did not go with hand + board.
I expect this Heuristic to hold true about 20% of the time.
When I deviated from the Heuristic it was for one of three reasons:
- Kill an important target or board clear.
- Make a big play.
- Grow the number of cards of a specific house in my hand in preparation for a big play.
This seems to show that the real depth of KeyForge happens when you deviate from the Heuristics. But sometimes there are many more options than there are apparent.
Spectating KarenB vs UltimoFox
It just happened that I opened TCO and saw a passworded game by KarenB. I knew there is a decent chance it was a leaderboard game (Go check it out, it’s cool), so I went to check the swindle discord for the password and found it. I immediately jumped in and since my recording software was open and I was alone I decided to commentate. I’ll excerpt the important part here.
My surprise prompted Tutenkharnage to analyze the play on the #radiant-truth channel of the Sanctumonius-Timeshapers Discord. The rest of this article is me paraphrasing their analysis.
Calling shadows is the hand + board play that results in drawing more cards and has the best potential of increasing your hand + board play the following turn. It does so with a delta of 5 (Opponent -1 and You +4) which is very significant. If you don’t call shadows now you’re almost certain to call it the following turn. The only case where you might not is drawing 2 good Sanctum cards, which has a low chance of happening.
However, looking over your opponent’s list, you can see a plethora of small creatures in both Logos and Shadows. Using your Pawn Sacrifice and Relentless Whispers now reduces your ability to deal with them later. Furthermore, the Shadows play eliminates one of the targets for Three Fates, making it less likely to find a good opportunity to play it later.
Your opponent just played three Dis cards and already opened with one. There is about a 36% chance your opponent has no Dis cards in hand and about a 46% chance they have just 1. This makes Three Fates + Control the Weak a free turn. Considering the makeup of your opponent’s deck, killing two of their biggest creatures with Three Fates seems ideal.
That said, there are a number of downsides to playing Dis. Killing your own Sergeant Zakiel with a captured Æmber is definitely not ideal. You’re almost certainly going to call Shadows next turn, but since you’re drawing only two cards, you have roughly a 47% chance to not draw any Shadows next turn, in which case your next Shadows turn will be no more impactful than it would be now.
Because you’re going to name Dis as the Control the Weak house, the likeliest next turn involves your opponent playing one or zero cards, you calling Shadows, and then you sacrificing Silvertooth anyway to damage one creature that you can probably destroy with Relentless Whispers. In other words, you’ve reset the board after your turn, except that you’ve used a lot of good answers to Professor Sutterkin and other small and troublesome creatures and given the board tempo to the opponent.
“Overall, I think calling Shadows is the better play, but I was glad Aurore had a different opinion because it led me to take a closer look at the numbers and think about the situation more broadly,” Tutenkharnage says.
Galaxy brain play?
After realizing the Shadows play after a Control the Weak into Dis will be suboptimal, Tutenkharnage asks “What if you called a different house with Control the Weak?” And then answers.
Going over your opponent’s list shows it has a lot of small Logos and Shadows creatures that are very susceptible to Pawn Sacrifice and Relentless Whispers:
Logos: 3x Archimedes, Pip Pip, Helper Bot, Professor Sutterkin, ZYX Researcher.
Shadows: Gamgee, 2x Ronnie Wristclocks, Umbra.
There’s no reason to think our opponent held back on playing Dis cards last turn, which means they ended the turn with three cards in some combination of Logos and Shadows, and they were drawing three more cards that were slightly more likely to be from those same two houses. So it’s very likely that the opponent has a hand full of Shadows or Logos cards, maybe even three or four from one house.
Control the Weak into Logos
Seven small creatures that die to the cards in our hand, two Pokes that won’t damage anything, two Brain Eaters, and Strange Gizmo. If we ever wanted to start messing with that set, now’s a good time to do it. In the worst-case scenario, our opponent plays something like two Archimedes next to each other, plus one or two more creatures. In that case, we can use Pawn Sacrifice to archive one Archimedes and a neighbor, then use Relentless Whispers to destroy the other. Also, note that even if we force a Dis turn, we’re very likely to call Shadows and waste our powder, unless we do something very suspect, such as “play Silvertooth, hold all the cards, and hope we still have something to use for Pawn Sacrifice,” at which point our opponent is going to come out with this board anyway. Only our tools for dealing with it will be in the discard pile.
Control the Weak into Shadows
Forcing a Shadows play means Ronnies have to steal 1A now and then work to steal 2A later. it also means Sucker Punch can’t be played and archived. Whistling Darts? No targets. Throwing Stars? No targets and no Æmber. We can destroy up to three creatures next turn.
Overall, forcing Logos with CtW in order to blow up the small nonsense might be an even better play than forcing Dis and getting the opponent to burn a turn, given our hand and the future board state.
It’s possible an even better play might be to call Shadows and make the same plays KarenB did, but sacrifice Zakiel instead of Silvertooth. Doing so will sacrifice a fraction of an Æmber, which represents the odds that Zakiel would have still been on the board the next time we called Sanctum. We just played Shadows and are less likely to come back to Shadows before we would call Sanctum, and since Silvertooth is quite fragile compared to Zakiel, we’re definitely taking a loss in that department. But the difference is that the board is now Skullion (P7) and Silvertooth (P2) instead of Skullion (P7) and Zakiel (P4), which means that we are going to be able to nuke any two P2 creatures that hit the board, rather than just one.
This means we can call Dis the following turn no matter what, destroy Skullion and any two P2 creatures of our choosing (unless a wild Brain Eater appears, in which case we get to destroy those two and an enemy P2 creature instead of those two and Zakiel). And follow up with our pick for Control the Weak.
This is player 2 turn 2, and the depth here is substantial. Imagine how many non-obvious plays we overlook when board states are full of complicated creatures and interactions. KeyForge is a very deep game. Make sure to explore it.