Video summary

The Difference Between Addicting & Annoying Roguelikes

Main summary

Key takeaways

Gaming

Storyline / Premise

  • “Blackout Jack” is framed as a roguelike that’s essentially a multi-hand blackjack run.
  • Each run is structured around milestones you must beat by earning enough progress chips (from hands that perform well against the dealer).
  • As the run progresses, you build an “engine” of cards and effects that helps you meet—and exceed milestone requirements more efficiently.

Gameplay Loop & Highlights

Core Loop

  • Play multiple blackjack hands at once.
  • Improve your odds by mixing/sacrificing bad hands to strengthen better hands.
  • After encounters/milestones, you gain access to upgrade options, including:
    • shops
    • boons/cards/effects
    • limited items

Engine-Building Components

  • Boons: relic/joker-style perks that change outcomes or rewards.
  • Card effects: permanent enchantments that upgrade your cards.
  • Elixirs: limited-use items that provide temporary advantage.

Types of “Engine Building” Actions

  • Type 1 actions: move you closer to victory (e.g., winning hands / getting 21s).
  • Type 2 actions: don’t win immediately, but make Type 1 actions stronger (e.g., shop purchases, adding cards, gaining effects).

What the Video Says Makes an Addicting Roguelike (vs. Annoying)

1) Avoid a “same-run every time” strategy (create real engine depth)

Problem identified (from early playtests)

  • The game became: “get 21s” as efficiently as possible.
  • Boons/effects boosted chips, but didn’t meaningfully change how players made decisions.

Fix

  • The dev enumerated more possible blackjack outcomes/actions, including (examples mentioned):
    • doubling down
    • hitting
    • splitting
    • standing with two cards
    • winning with a 16
    • making the dealer bust
    • building hands limited to certain suits (e.g., only diamonds)
  • Many actions gained items that enable different viable strategies, such as:
    • a playstyle that rewards purposefully busting (turning a “bad” action into a strong build goal)
    • strategies focused on:
      • winning with low values
      • building large life-token income for sacrifice-based play
      • streak-based seat pressure / sabotage themes
      • stash/flexibility-focused builds
  • Result: builds feel like crafted engines, not one “best play.”

2) Synergies must drive the “this is a build” feeling

Problem identified

  • Shops became “pick the always-best item,” making choices feel solved.

Fix

  • Make power come from synergies, not a single always-OP pick.
  • The dev also notes emergent/hidden synergies can appear—even unintended—such as a described near-infinite loop:
    • Wild split → split any hand
    • Blood offering → sacrifice split for life tokens
    • Soul tap → gain mana by sacrificing
  • This kind of loop was later tweaked/balanced.

3) Increase “build upgrade moments” (“bum/bombs”) so players can adapt

Key idea

  • Upgrade decision opportunities are crucial to the sense that the run is yours.

Problem identified

  • Early version had ~10 upgrade moments per run, which felt too low.
  • Players needed more chances to make brick-by-brick decisions and recover from bad RNG.

Fixes

  • Added more upgrade moments between rounds by:
    • introducing a dice rolling mechanic: instead of playing blackjack each round, spend life tokens to roll for additional upgrades
    • this roughly tripled the number of upgrade moments
  • Made the shop more flexible:
    • not only at milestones, but between rounds when money is earned mid-milestone

Variable upgrade moments (depth + strategy)

  • Variable “bum” amounts add strategic layering (similar to FTL encouraging extra visits to improve odds).
  • Ways to make “bum counts” variable:
    • shop visitability during milestones (if players can earn money mid-milestone)
    • rolling the shop for money (classic variable reward mechanic)
    • the dice mechanic itself: rolling is risky because it costs tokens and reduces remaining rounds—an intentional tradeoff

4) Give players a stronger internal locus of control (choice-maxing)

Psychological goal

  • Roguelikes should make players feel they won through strategy, not external luck.

Fix: increase meaningful choices

  • Reframe random outcomes into player selections, e.g.:
    • a round upgrade that used to apply to one random card now shows five random cards and lets you choose
    • seat upgrades that used to pick a random seat now let you choose the seat

5) Make restarting not feel awful: bend the difficulty curve + add early variety

Problem

  • Losing resets the run. If early difficulty is too steep—or early runs feel identical—restarts become punishing and boring.

Difficulty-curve solution (“bend the fun channel”)

  • The dev argues for:
    1. easy early game for new players
    2. but early game as the hardest strategic phase (experts optimize it)
  • Analogies:
    • FTL: early sector isn’t hard, but doing well by visiting more encounters and staying unscathed is hard
    • Slay the Spire: optional elite battles offer better rewards but are harder choices suited to skilled players

Blackout Jack-specific addition

  • Two types of round bonuses:
    • early bonuses: reward reaching the milestone with time/rounds to spare
    • overshoot bonuses: reward exceeding the milestone significantly
  • This incentivizes strong early play, improving later odds because better performance funds more dice/bum rolls.

Early variety solution: starting classes

  • To prevent openings from feeling identical, the game unlocks multiple starting classes, each with a different immediate playstyle:
    • Default: increases draw into stash
    • Dog class: reduces sacrifice cost
    • “Cutp” class: increases dice gained when you roll

6) Reduce the “RNG shaft” feeling (risk without helplessness)

Dilemma

  • Roguelikes need randomness for replayability.
  • But randomness can sometimes feel unfair (e.g., dealer hitting perfect hands).

Mitigations mentioned

  • Provide options to prevent bad RNG—at a cost, such as:
    • items/cards that reduce the dealer’s ability to reach 21 / increase dealer bust chances
    • players can buy more dice with life tokens, so a bad roll is partly tied to a choice to invest (or not)
  • Improve adaptability when things go wrong:
    • if the dealer gets a natural blackjack (21 on the first two cards), the dealer reveals it immediately
    • the player then gets actionable information, e.g.:
      • use elixirs to sabotage
      • use boons to improve tie outcomes
      • sacrifice hands to farm cards for the next round

Key Tips / Design Takeaways Emphasized

  • Create multiple viable strategies by designing items around many different actions, not just one baseline path.
  • Ensure shop decisions aren’t solved by enforcing synergy-driven power and preventing a single always-best item.
  • Add enough—and variable—upgrade decision moments (“bum(s)/bombs”) so players can adapt and keep agency.
  • Use choice-maxing (offer selection instead of random auto-targeting) to strengthen internal locus of control.
  • Bend the difficulty curve so:
    • early game is accessible,
    • but expert success comes from strategic optimization early.
  • Add early-run variety (e.g., starting classes) so restarts feel new.
  • When bad RNG happens, give players:
    • ways to prevent it (through investment),
    • and ways to adapt once it happens (through information and choices).

Gamers / Sources Featured (Mentioned)

  • Reddit (a “random person on Reddit 5 years ago” cited for the engine-builder definition)
  • Slay the Spire
  • FTL
  • Balatro
  • Clover Pit (referenced as a comparison for how certain mechanics resemble it)
  • Breaking Bad (quote referenced: “I am not in danger… I am the danger.”)
  • Near: Automata (mentioned as a counterexample for painful restart experience)
  • Steam Summer Sale (context for acquiring Near: Automata)

Original video