Astrea: Six-Sided Oracles

I have a hard rule about roguelikes: if I can solve them on autopilot within three runs, they go in the bin. Astrea: Six-Sided Oracles has not gone in the bin. The core swap - cards replaced by physical dice you roll each turn - sounds like a surface-level gimmick until you hit your first elite fight and realize that every face of every die in your pool is a probability you chose, and a bad choice three rooms ago is now killing you. That accountability loop is exactly what keeps the genre interesting, and Little Leo Games, a Brazilian indie studio, built it into every layer of the design. The mechanical centrepiece is the Purification vs. Corruption duality. Purification damages enemies and heals you; Corruption does the reverse. The tension comes from dice that can roll either result, and a health bar that doubles as a resource - you can deliberately absorb Corruption to trigger powerful Oracle-specific abilities, but tip too far and you lose the run. On top of that, each of the six playable Oracles has a radically different kit. Moonie (an owl) is the safe starter who can flip corrupt faces to pure. Cellarius, a hammerhead shark monk, Purifies enemies while intentionally self-corrupting, sometimes detonating waves of purifying energy across the whole field. Hevelius enters every fight with two Sentinel companions - upgradeable mechanical constructs with their own dice rolls - but pays for it with a thinner health pool. Later Oracles, including the final character Orion, require completing runs with all prior characters first, giving the unlock ladder a genuine sense of earned progression. Spending time in the community discussions, it is clear that Sothis and Orion polarize players the most: Sothis rewards exponential ramp strategies that feel useless on floor one and broken by floor three, while Orion demands extremely specific blessing synergies to function at all. The Blessing system is where the build-crafting really opens up. Star Blessings offer reliable passive bonuses; Black Hole Blessings offer dramatically stronger effects with meaningful drawbacks, things like drawing extra corrupt dice each turn or capping your hand size. Choosing a Black Hole Blessing is the kind of decision I want to think about for thirty seconds, which is the right pace for this genre. Dice themselves are categorized as Safe (mostly positive faces), Balanced, or Risky (two or three positive faces but enormous upside when they land). Community debate on the forums splits between players who argue that optimized endgame pools should be lean and risky-heavy - sometimes as few as eight forged dice - and those who have won convincingly with safe-die-heavy decks using Virtue-refresh combos. That breadth of viable approach is a good sign for long-term replay. The anomaly difficulty system, which functions similarly to Slay the Spire's ascension ladder, provides sixteen levels of increasing challenge with accelerated XP gain, keeping high-hour players invested. The legitimate criticisms are real, though. The UI struggles to communicate stacked status effects clearly, and early runs often end to damage the player had no clean way to anticipate. The onboarding assumes you will bounce off the first few hours and come back; players who need a tutorial to fully hold their hand will find the first few sessions disorienting. A speed scale option buried in settings is essentially mandatory for comfortable pacing - that should be front-and-center. There is also a mild RNG frustration at high anomaly levels, where a dice pool that should succeed on four of six faces will somehow hit the bad face five turns in a row. That is variance working exactly as probability allows, but it does not feel great in the moment. Post-launch, the developer has been actively pushing balance patches through beta branches, reworking specific Oracle starting dice, nerfing overtuned blessings, and iterating on anomaly scaling - the update cadence suggests genuine long-term investment in polish. For strategy and roguelike players who clock measured, build-focused runs: this is a serious contender. The depth-per-hour ratio is among the best in the dice-builder sub-genre, the six Oracles justify multiple complete playthroughs on their own, and the combination of over 170 Blessings and hundreds of available dice means the problem space never fully collapses. If you bounced off it once, try adjusting the speed setting and committing to Moonie for three full runs before deciding. The game does not show you everything it is in the first sitting.

Diego · Scout Team

Strategy & simulation

Be the first to comment on Astrea: Six-Sided Oracles.