Dwarf Eats Mountain 100% 速通路線：從準備局到第 100 座山

This route assumes permanent Prestige progress, starting resources, and broad building access. It is not a first-playthrough promise. Use the splits as decision gates rather than exact timestamps because artifacts and version balance create variance.

1. Define the category before starting

Record the save state that makes the run possible: starting gold and Mithril, unlocked buildings, retained artifacts, Tier 6 bonuses, and allowed speed settings. Without a fixed baseline, a completion time cannot be compared fairly.

The route optimizes three sources of lost time:

• waiting for enough economy to buy the next damage breakpoint;
• arriving at a boss without the required uptime or burst;
• reaching the final gate while still missing a unique artifact.

2. Opening: activate economy and research immediately

Use starting resources to establish the main early damage engine, the research required to scale it, and the exploration system that begins time-gated artifact rolls. Avoid buying every building simply because it is available.

The observed route opened with Lasers, a small Scientist group, and early exploration. Its purpose was to turn damage into gems while the artifact clock ran in parallel.

Exit check

• Ore does not remain unpaid on the ground.
• The main weapon has a clear next breakpoint.
• Artifact and key collection has already begun.
• No housing slot is occupied by a unit without a route purpose.

3. Early–mid game: buy breakpoints, not comfort

Add Scientists until the next marginal purchase no longer unlocks a meaningful Laser or research gain. Add Priests, Stone Callers, or equivalent support only when they start the Faith or calamity-proof economy needed later.

This is also where a speedrun differs from completion play: a comfortable defensive purchase may be skipped if it does not reduce expected time, while a reliable logistics upgrade is kept because a single stalled delivery line can lose an entire split.

4. Mid game: route the mandatory systems in parallel

Between boss gates, four tasks should progress together:

1. deepen the primary damage engine;
2. collect keys and scheduled unique rewards;
3. establish Faith and underground returns;
4. reserve forge resources for missing pool items.

If one task is left until the others finish, it becomes forced waiting later.

5. Pivot when the scaling leader changes

The recorded run used Lasers early, then gained stronger Ballista and Flame Lord scaling later. The exact winner can change, but the pivot rule is stable:

• compare real clear time at 1× speed;
• include the economic value of each unit, not only displayed DPS;
• move new resources to the engine with the best future multiplier;
• keep the old engine if it still generates gems, protects Dens, or supplies another route function.

A pivot is a reallocation of new investment, not necessarily demolition of the old build.

6. Forge and artifact routing

List every endgame item and mark whether it is guaranteed, boss-locked, expedition-based, or Forge-only. Delay mass forging until the route has any refund or efficiency effect it expects, but do not delay so long that the final gate becomes a lottery.

Set a resource floor: Mithril and Souls below that number are reserved for missing mandatory items. Only the surplus can fund optional damage.

7. Final push: remove decisions from the run

By the last phase, the route should be a checklist:

• verify the full artifact set before the gate;
• move all new damage investment to the current scaling leader;
• keep logistics ahead of the larger loot volume;
• use 1× speed for boss validation even if fast-forward is used elsewhere;
• stop buying upgrades whose payback time exceeds the remaining run.

The fastest late-game purchase is often no purchase at all.

8. Review by lost time

After the run, do not only compare total time. Record each delay under one cause:

Cause	Example	Route fix
Economy	Waiting for housing or Mithril	Move the income breakpoint earlier
Execution	Missed ability or bad positioning	Add a pre-boss checklist
Routing	Missing key or Forge item	Start the source earlier or reserve more
Build	Damage engine pivoted too late	Define a measurable pivot trigger
Version	Values or gates changed	Update the route baseline

Optimize the largest repeatable loss first. Random artifact variance should be noted separately from route mistakes.