# Growth plan — averages, not wishes

Everything below is computed from measured numbers, not assumed ones. Where a number is
extrapolated it says so. Source: `bt_events_fee0.0_line_touch3_t0_fu0.000012.csv`
(2026-10-15 → 10-22, 13,107,223 ticks, ZEC, 20x, 0% fees, real spread charged).

## 1. The measured averages

| Measure | Value |
|---|---|
| Ticks replayed | 13,107,223 (7.6 days) |
| 1hr lines printed | 1,398 |
| Cycles booked | 3,309 |
| Cycles that paid | 56 (**1.7%**) |
| Cycles flat | 3,249 (98.2%) |
| Cycles that lost | 4 (0.12%) |
| Bricks banked | **673** |
| Bricks per booked cycle | 0.20 (mean), 0 (median) |
| Longest 1hr run | 131 bricks |
| Runs ≥50 bricks | 6 |
| Gross before costs | $12.51 |
| Spread paid | $4.63 (**37%** of gross) |
| Funding paid | $0.017 (0.1% of gross) |
| Net | $7.86 on $1 start |
| Max drawdown | 17.7% |
| **Average daily compounding** | **+31%** ($1 → $6.52 in 7.6 days) |

The shape of the edge, in one line: **almost everything is flat, a few cycles carry the
month, and one 131-brick run pays for dozens of empty hours.**

## 2. Projection from that average (extrapolation, not a promise)

At +31%/day compounding, before any capacity limit:

| From | To $1,000 | To $100k | To $10M | To $800M |
|---|---|---|---|---|
| $10 | 15 days | 31 days | 45 days | 61 days |
| $1,000 | — | 16 days | 30 days | 45 days |

Days are calendar days of continuous trading. **This curve breaks long before $800M** —
section 3 explains where.

## 3. Where the curve actually bends

| Capital | Binding constraint | What breaks first |
|---|---|---|
| up to ~$10k | nothing structural | spread, fill quality |
| ~$10k–$500k | Lighter ZEC depth | your own orders move the book |
| ~$500k–$5M | venue concentration | one exchange holding your whole stack |
| ~$5M–$50M | position size vs open interest | 1% of OI is a few hundred thousand |
| $50M+ | market impact | 0.5% of daily volume per day, per the sector rule |

The 800M plan's own note is right: each sector absorbs ~0.5% of its daily volume without
moving price, so the last $300M alone takes about 23 trading days to place — after you
already have it. ZEC alone cannot absorb it; it needs BTC/ETH and several venues.

## 4. Milestones and what changes at each

1. **$10 → $1,000 on Lighter testnet, then mainnet.** Same rules, $10 risk per trade.
   Withdraw 60% of every profitable cycle to your own wallet.
2. **$1,000 → $25,000.** Keep ≤30% on Lighter. Move winnings to a second venue.
3. **$25,000 → $250,000.** Rest limits only (0% fees both sides), size capped at 1% of
   visible depth. Start measuring your own slippage against the model.
4. **$250,000 → $5M.** Multi-venue: Lighter + Hyperliquid + one more. Correlated
   positions across venues are one risk, not three.
5. **$5M+.** Slice by percentage of ADV: ≤0.5%/day per market. This, not the strategy,
   sets the growth rate from here on.

## 5. The five rules that protect the compounding

1. **Withdraw 60% of every profitable cycle.** Money on the exchange is at risk; money in
   your wallet is yours.
2. **Never more than 30% of total capital on one venue.**
3. **One position per market**, always with an exchange-side stop 2 boxes beyond the line.
4. **Daily loss cap** — the runner already enforces 25% per market per day.
5. **Kill switch** — `STOP` file or `/stop` on Telegram flattens everything and exits.

## 6. What would make this document wrong

- A 7.6-day sample is short. One month of chop could produce a flat stretch where
  `median -0.08 bricks per cycle` dominates and the +31%/day average collapses.
- The 37% spread drag assumes market orders. Resting limits remove it — and add
  unfilled entries, which this replay does not model.
- Fill probability at real size is unmeasured. Treat every number above as an upper bound
  until it is measured on the venue.