Krodha Combat Variant Lab — demo order

25 total playable prototypes. This page is a presentation order, not canon. It is intentionally ordered from weakest → strongest model so you can demo the search as a rising curve.

Docs: the original documented set is 19 (`V01–V19`). `V20–V25` are expansion experiments added to hit the 25-prototype target.

Demo notes

Weakest first
Arithmetic burden, lower legibility, or weaker design signal.
Strongest last
Best mix of intuitiveness, element load-bearing, and player-read clarity.
Selected
Top 10 from `01-SLATE.md`.
Overflow
Built from discarded candidates / expansion wave.

Core docs

00-CHARTER.md — hard constraints / grounding 01-SLATE.md — selected 10 + original 19-candidate scoring 02-OVERFLOW-SLATE.md — V11–V25 build tracking

Jump to strongest five

#21 — V05 Sangam #22 — V10 The Written Die #23 — V01 House Rules #24 — V09 Their Dice #25 — V02 Four Braziers
1

V16 — The Churn

overflowweakest tier

Queued-element / next-turn transform model. Interesting as a thought experiment, but the planning burden is high and the board read is thin.

Open V16
2

V11 — Forge

overflowcompute-heavy

Arithmetic rung-shift version of House Rules. Teaches why "do math before reading the board" is the wrong direction even when the outputs are coherent.

Open V11
3

V14 — Kalari Stance

overflow

Enemy-facing stance read. Stronger than Churn/Forge, but still puts too much meaning outside the player's authored board.

Open V14
4

V19 — Fuse

overflow

Enemy fuse track. Tense, but very binary: much of the experience collapses into "did I overheat the bomb?"

Open V19
5

V15 — Engraving

overflow

Face-state persistence variant. Useful contrast piece, but denser and fussier than The Written Die.

Open V15
6

V18 — The Walker

overflow

Moving focus marker plus trail logic. Has real motion, but asks the player to read both destination and aftermath at once.

Open V18
7

V17 — Declare

overflow

Call-your-shot model. Clean enough to explain, but it narrows the board and makes misses feel like negation rather than adaptation.

Open V17
8

V13 — Manthan

overflow

Current element vs last-turn element. Better than Churn, but still less agentic than using a board-authored second operand.

Open V13
9

V12 — Weather Die

overflow

Whole-board weather swing. Exciting, but it tramples the placement layer more than the selected slate wanted.

Open V12
10

V24 — Pressure Dial

new

Shared pressure band. Readable and presentable, but it still sits a bit above the board instead of inside it.

Open V24
11

V21 — Counterweight

new

Left/right board lean. More legible than Pressure Dial, with a stronger spatial read, but still less vivid than the best selected variants.

Open V21
12

V23 — Tributary

new

Board flow direction. Good motion and nice sequencing texture, though it competes with placement for attention.

Open V23
13

V22 — Forecast

new

Number-memory omen system. Fairly intuitive, cleaner than the element-memory versions, but still a touch abstract.

Open V22
14

V25 — Gatehouse

new

Open-gate patterns can fire multiple slots at once. Strong board readability, slightly busier than the top models.

Open V25
15

V20 — Blank Bargain

new

Best of the new expansion wave: blanks become a real economy without losing the core slot/dice read.

Open V20
16

V06 — Tide Price

selected

Late selected model. Clear and promising, but not as instantly sticky as the very top finalists.

Open V06
17

V03 — Twin Wells

selected

Distinct economy identity. Strong search artifact, though slightly more bookkeeping-heavy than the very best demo pieces.

Open V03
18

V07 — One Wick

selected

Track-based and flavorful. Better than the weaker overflow set, but still not the cleanest high-level pitch.

Open V07
19

V08 — Fuse Line

selected

Reaction model with a strong identity. Good middle-to-late demo piece once the audience already understands the search space.

Open V08
20

V04 — Painted Ground

selected

Persistent anchor state and resonance. Very presentable, very game-like, and visually easy to talk through.

Open V04
21

V05 — Sangam

selectedstrong tier

Excellent reaction-wheel model. Feels authored, legible, and richer than Manthan without losing clarity.

Open V05
22

V10 — The Written Die

selectedstrong tier

One of the best wildcard models: distinct, memorable, and much cleaner than Engraving.

Open V10
23

V01 — House Rules

selectedstrong tier

Closest to the live core while still making elements truly load-bearing. A superb baseline for the whole search.

Open V01
24

V09 — Their Dice

selectedstrong tier

Extremely sharp late-demo piece. Strong player read, enemy contrast, and a good sense of "this could be a real game direction."

Open V09
25

V02 — Four Braziers

selectedstrongest tier

The strongest presentation model in this ordering: extremely readable, obviously elemental, and immediately demoable to another human.

Open V02