Why You're Taking Too Much Damage (10,000 Simulated Combats Prove It)

Storyline / Context

• The video focuses on Slay the Spire 2, specifically early combats against common enemies (using combat simulators).
• There’s no narrative “story” plot discussed. Instead, the “story” is:
  • how fights escalate, and
  • how player decision-making changes the outcome as difficulty (Ascension) rises.

Gameplay Concepts Tested

The creator compares four combat decision “strategies/algorithms” by running simulated fights and tracking how much damage the player takes:

• Block first

  • Block until you take 0 damage, then play attacks.
  • Common among newer players.

• Minimize over block

  • Example: if the enemy attacks for 11, block for 10, take the 1, and move on rather than blocking extra.

• Attack first

  • Play attacks until you run out, then defend.
  • Generally worse early, but can be optimal in some specific enemy patterns.

• MC rollout (Monte Carlo rollout)

  • A look-ahead planning method.
  • It simulates likely future draws and enemy actions (not perfect information).
  • Looks roughly 3–4 turns ahead, choosing actions based on expected outcomes across the future—not just the current turn.

Data visualization note

• Damage in the game is quantized/discrete (e.g., taking 13/18 rather than 14/15).
• Charts are dequantized for readability.
• This is meant to show overall trends, not exact probabilities.

Key Findings by Enemy (Gameplay Highlights + What the Data Shows)

1) Fuzzy Worm / Worm Crawler (simpler fight)

• Has repeated attacks (two-hit pattern with strength scaling).

Low Ascension (around A8):

• MC rollout ≈ Min over block (planning ahead adds little).
• Best approach: kill efficiently and don’t stall.
• Block first becomes clearly worse
  • A long tail appears where overly passive play leads to taking very large hits later as enemy strength ramps.
• Attack first is generally poor here.

Higher Ascensions (A9–A10):

• The cost of Block first becomes even harsher.
• “Optimal-ish” play saves only a few points versus basic strategies, but those few points matter more at higher difficulty.

2) The Nib it (mid-complex fight, scaling pressure)

• Pattern: a big hit, then a hit-and-block, then it gains strength; repeats.

A8:

• Planning ahead helps, but not dramatically.
• Minimize over block is sufficient.
• Full “block everything” is unreliable.

A9–A10:

• Block first averages much higher damage.
• MC rollout provides a larger, measurable benefit by planning defenses and attacks over upcoming turns.

Nuance:

• The “optimal” referenced is not mathematically perfect—it’s a noisy approximation humans can realistically approach.

3) Elite fight example: “Snecko King Colony”

• Tested with a simplified “elite-ready” starter deck:
  • 2 Ultimate Strikes + 2 Ultimate Defend
• Key mechanic emphasized:
  • You can only deal up to 15 damage per turn.

Major takeaway:

• MC rollout delivers a huge advantage, saving substantial damage by planning turn sequences.

Additional observation:

• Full blocking can outperform “minimize over block” for this specific elite (not fully explained, possibly due to frequent large attack values).

Community misconception addressed:

• People often say elites are impossible.
• The video argues the real issue is how you play, not enemy RNG or your starting cards.

4) “Birdonus” (burden of pushing damage vs defending too long)

• Has two attack types with strength increasing each turn.

Simple best strategy:

• Attack first
• Generally deals less damage taken than “block all the time,” because pressure/offense matters.

A8 → A9:

• “Just play what’s in hand” becomes extremely costly on Ascension 9.
• MC rollout keeps damage far lower, framed as:
  • basic play being unbeatable, versus
  • rollout making it manageable.

Key Tips / Strategies the Video Recommends (Practical)

• Don’t just react with this turn’s cards—plan 1–2 turns ahead.
• Use a damage plan before committing plays:
  • Ask: Can I kill this turn / next turn / following turn?
  • Consider what the enemy will attack on upcoming turns, not only what they’re doing now.
• Slow down intentionally
  • Take extra seconds to think instead of playing immediately.
  • The video presents this as the biggest real win-rate factor revealed by simulations.
• If you can’t be optimal:
  • Methodical planning beats fast, autopilot play.
  • Potions provide extra safety (example mentioned: mitigating damage like taking ~10 instead of being overwhelmed).

“So what?” Takeaway (Creator’s Main Message)

• Difficulty spikes (especially around Ascension 9) aren’t explained by “bad cards.”
• They’re driven more by suboptimal turn planning.
• Elite fights are often not impossible—players lose by not planning future consequences, causing basic strategies to suddenly take about double the damage.

Gamers / Sources Featured

• No specific other gamers, streamers, or sources are named in the subtitles.
• The video references general resources about stochastic decision-making algorithms, without citing a particular author or platform.