STACK-X0 — Polymarket Crypto Up/Down Bot

Status update. Current market conditions are unfavorable for this bot. An updated strategy is in progress — see the recording below for context from @stack-x0 (0x9a07c6583fb9defd31a102add491d35621c404e1):

Watch full update (MP4)

Automated trading engine for Polymarket crypto Up/Down markets (BTC, ETH, SOL, XRP — 5m and 15m windows). Strategy core:

One side. One entry band. Hold to redeem. Never sell. Never hedge.

For every active slot the bot:

1. Reads direction from CEX lead vs slot open (spot − open gap)
2. Identifies the favorite side (book prices it ~52–57¢)
3. If favorite ask ∈ [50¢, 70¢] and conviction is OK → taker BUY ~$15–$60 (~50–100 shares)
4. Holds to resolution → redeem $1 if correct, $0 if wrong
5. Never sells, never buys the opposite side

Built on @stack-x0 and wallet 0x9a07c6583fb9defd31a102add491d35621c404e1.

Access: This bot is intended for active membership owners. A free 1-day trial is available for testing; after the trial expires, contact me on Telegram at @StackX0 to activate membership access.

Disclaimer: Observed public patterns + an independent reimplementation. Not financial advice. Past results do not guarantee future performance. Start with simulation and micro-size live clips.

Table of contents

1. Strategy in one page
2. Why this can be profitable
3. The math — edge, fees, break-even
4. Conviction and the entry band
5. What we do not do
6. What kills the edge
7. Visual walkthrough (screenshots)
8. How the bot implements it
9. Quick start
10. Configuration
11. Simulation and validation
12. Production
13. Architecture
14. Further reading

Strategy in one page

For each active crypto up/down slot (BTC / ETH / SOL / XRP, 5m or 15m):

  1. Read direction from CEX lead vs slot open (spot vs open gap)
  2. Identify the FAVORITE side (book prices it 52–57¢)
  3. If favorite ask ∈ [0.50, 0.70] and conviction OK:
       → TAKER BUY ~$15–$60 (~50–100 shares)
  4. Hold to resolution → redeem $1 if correct, $0 if wrong
  5. Never sell, never hedge opposite side

flowchart TB
  slot[Active 5m or 15m slot]
  cex[CEX spot vs slot open]
  fav[Favorite side ask 52-57c]
  band{Ask in 50-70c and conviction OK?}
  buy[Taker BUY 15-60 USD]
  hold[Hold — no sell no hedge]
  redeem[Redeem at resolution]
  slot --> cex --> fav --> band
  band -->|yes| buy --> hold --> redeem
  band -->|no| wait[Skip or wait next tick]

Stack-X0 profile context

The @stack-x0 profile shows a live wallet around $6.9k portfolio / $6.8k cash and a recent activity stream dominated by mid-band crypto Up/Down buys. The visible rows include entries such as ETH Down 46.0¢ / 53.6¢ and BTC Down 49.0¢, matching the “favorite/mid-band, hold-to-redeem” style this README documents.

Why this can be profitable

This is not endgame noise farming (sweeping 94–99¢) and not lottery spray (buying 3–30¢ losers). It is a mid-band favorite trade:

• You buy when the book and CEX already agree on a mild favorite.
• You pay roughly 50–70¢ per share for a claim that pays $1 if you are right.
• Gross payoff on a winning entry at 55¢ is about +82% on capital deployed (before fees).
• Gross payoff on a losing entry is −100% of that clip.

Profitability comes from asymmetric dollars per win vs loss at moderate prices, combined with repeated slots across assets:

Because wins pay more dollars than losses lose on the same size only if win rate stays high enough after fees — otherwise the strategy loses. Mid-band favorites aim for that: the CEX gap + book confirmation filters coin-flips; you skip both razor-thin 99¢ clips and cheap underdog lottery tickets.

flowchart LR
  subgraph edge [Where edge comes from]
    A[CEX leads Polymarket]
    B[Book still mid-priced 50-70c]
    C[Favorite not fully priced in]
  end
  subgraph pnl [PnL]
    W[Win: redeem near 2x at 50c]
    L[Lose: lose full premium]
  end
  A --> C
  B --> C
  C --> W
  C --> L

Core claim (expectancy):

If, after fees, your long-run win rate on these mid-band favorites exceeds the break-even rate for your average entry price, the strategy has positive expectancy. Volume (many BTC/ETH/SOL/XRP slots) then compounds that edge.

Winning-history evidence

The history screenshot shows the exact mechanics: mid-band buys around 48–55¢ and later green Redeem rows. Examples visible in the screenshot include XRP and ETH redemptions around +$78.79, +$96.25, +$87.29, +$28.42, and +$127.86. This is the core cash-flow path: buy one side, wait, redeem if correct.

The math — edge, fees, break-even

Gross vs net after taker fees

FOK buys are takers. Crypto 5m/15m fee rate ≈ 0.072:

feeShares = shares × 0.072 × price × (1 − price)
netEdgePerShare = 1 − price − (0.072 × price × (1 − price))

*Rough (premium risk / (premium + net win)); actual break-even rises slightly after fee share haircuts on fills. Use sim logs to calibrate your live filled prices.

Read this carefully: at 70¢ you need a much higher hit rate than at 50¢. That is why the band is capped at 0.70 — above that, edges shrink and break-even win rates climb into “almost resolved” territory where you are paying too much for residual upside.

Worked example — one slot, one clip

CEX: BTC spot $98,420   Slot open (price to beat): $98,380
Gap: +$40 → lean UP

Book: UP best ask 0.54   DOWN best ask 0.47
Favorite: UP (~54¢ in the 52–57¢ “mild favorite” zone)
Ask in [0.50, 0.70]: yes
Conviction: gap + book agreement OK

→ TAKER BUY 80 shares UP @ 0.54  (~$43.20 notional)

If UP wins:  redeem ~$80  − fees  → ~+$35 profit
If UP loses: redeem $0            → ~−$43 loss
Never BUY DOWN. Never SELL UP.

Why size is $15–$60 (not $2 micro-clips)

This strategy sizes enough to make mid-band edge matter, but small enough that one wrong slot does not sink the wallet:

~50–100 shares at mid-band prices lands in this notional range by design.

Profile / analytics

Use the profile view to track wallet health: portfolio, cash available, one-day PnL, and the most recent unresolved buys. For this strategy, the most important KPI is not trade count — it is hit rate by average filled price.

Conviction and the entry band

Direction (CEX lead)

gap = spot − openPrice
gap > 0  → candidate UP
gap < 0  → candidate DOWN

Optional filter: require |gap| ≥ STACK_X0_MIN_GAP_USD so flat open ≈ coin-flip slots are skipped.

Favorite identification (52–57¢)

The favorite is the side the book already leans toward — typically mid ~52–57¢ (not 90¢ endgame, not 20¢ lottery). That band means:

• The market is not undecided coin-flip at 50/50 forever.
• The move is not fully baked into a 90¢+ ask yet (still room to $1).

Entry gate [50¢, 70¢]

One entry philosophy

• One favorite side per slot once conviction fires.
• No averaging into the opposite side.
• No hedge. If you are wrong, you eat the premium — that is the cost of the edge model.

Recent buy pattern from Stack-X0

Visible buys include BTC Down 50.0¢, XRP Down 50.9¢, ETH Down 50.8¢, SOL Up 55.5¢, and XRP Down 48.1¢. These are not 99¢ convergence clips or 10¢ lottery tickets — they are mid-band directional entries.

What we do not do

❌ SELL ...
❌ BUY underdog "just in case"
❌ BUY both UP and DOWN in the same slot
✅ BUY favorite once in [50¢, 70¢], hold, redeem

What kills the edge

Visual walkthrough (screenshots)

The README uses the two screenshots currently available under docs/screenshots/:

Profile screenshot

Winning-history screenshot

Together these two screenshots show the profile-level bankroll context and the trade-level cash-flow pattern: mid-band buys followed by redeem rows when the selected side resolves correctly.

How the bot implements it

Intended tick loop:

1. openPrice + CEX spot → gap → candidate favorite
2. Read favorite best ask
3. If ask ∈ [0.50, 0.70] and conviction OK and under market cap:
     TAKER BUY ~$15–$60 (~50–100 sh)
4. Stop further opposite-side logic (none exists)
5. Slot end → redeem winners

Note: The user-facing strategy name is stack-x0, and this README is centered on the Stack-X0 wallet 0x9a07c6583fb9defd31a102add491d35621c404e1.

Quick start

Prerequisites

• Node.js 18+ (or Bun)
• Production: Polygon wallet with pUSD (docs/MIGRATE_V2.md)

Install

cd stack-x0
npm install          # or: bun install
cp .env.example .env

Simulation

npm run stack-x0:sim
# npx tsx index.ts --strategy stack-x0 --slot-offset 1 --rounds 20 --always-log

Raise MAX_SESSION_LOSS (e.g. 200) for longer runs.

Multi-asset fleet

npx tsx scripts/run-stack-x0.ts
# optional: --window=15m   --prod

Production

# .env: PRIVATE_KEY, POLY_FUNDER_ADDRESS, BUILDER_*, FORCE_PROD=true
npm run stack-x0:prod
npx tsx scripts/redeem.ts   # backup redeem

Configuration

Engine

Strategy knobs (align with this README)

See .env.example and docs/STACK-X0.md for the current Stack-X0 env list.

Simulation and validation

npx tsx scripts/chart.ts logs/early-bird-btc-updown-5m-*.log --open

Checklist

• Direction matches CEX gap sign
• Entries only on one side (favorite)
• Ask prices in 50–70¢ (often clustered mid 52–57¢)
• Notionals roughly $15–$60 / ~50–100 shares
• Zero SELL lines
• Zero opposite-side buys in the same slot
• Winners redeem; losers go to zero

Production

1. Fund pUSD on the proxy wallet.
2. Sim ≥10 rounds; confirm mid-band single-side entries.
3. Set MAX_SESSION_LOSS and per-slot cap for bankroll.
4. Run fleet: npm run stack-x0:prod.
5. Monitor hit rate at filled price, average entry, and session PnL.
6. Run scripts/redeem.ts periodically as backup.

Architecture

scripts/run-stack-x0.ts
  ├── MARKET_ASSET=btc  → EarlyBird + stack-x0
  ├── MARKET_ASSET=eth  → EarlyBird + stack-x0
  ├── MARKET_ASSET=sol  → EarlyBird + stack-x0
  └── MARKET_ASSET=xrp  → EarlyBird + stack-x0
            │
            ▼
  MarketLifecycle → direction + favorite band → FOK buy
                 → hold to resolve → redeemPositions

Key files

engine/strategy/stack-x0.ts              # Strategy entrypoint alias
engine/strategy/lib/winner-inference.ts  # CEX spot vs open → UP/DOWN
engine/strategy/lib/fees.ts              # Taker fee / net edge helpers
engine/strategy/lib/inventory.ts         # Per-slot spend caps
scripts/run-stack-x0.ts                  # Multi-asset launcher

Further reading