ExamplesFor players

Pick your next deck

What is winning in the recorded corpus, weighted by results rather than by how often you see it on a stream — with the sample sizes attached.

A real exchange, recorded 2026-09-10 as King Thing of POAP KINGS: the tools under the window are the ones that answered, and every number came back from them.

Tower Sage your agent elixir-mcp connected
What should I play next season? I'm bored of my Witch deck.

Honest answer first: the corpus can tell you what is winning, not what will win for you. Over the last 28 days — 9,938 decided Ladder battles on record — the decks with the best win rates are all played by one pilot each, 60–65% over 30–68 games, so they say more about that player than about the deck.

The widely played decks tell the opposite story: the most common Ladder deck in the corpus, X-Bow / Tesla / Archers (42 different players), wins only 42%. Popular is not the same as good.

Two decks are worth a look because they overlap what you already play:

  • Witch / Skeleton King / Inferno Dragon / Mega Knight / Royal Hogs / Baby Dragon / Arrows / Ronin — a POAP KINGS clanmate is 25–2 with it this month, and six of its eight cards are in your rotation already.
  • Mortar / Golden Knight / Electro Dragon — your own second deck, 11–2. You may not need a new deck; you may need to play the one you are winning with.

Sample sizes are small everywhere here. Treat this as where to look, not a ranking.

What it reads
Meta decks by segment: the corpus, your clan, or one player Win rates from recorded battles, shrunk toward the corpus mean Your own deck history
To set this up
Request access
More like this