About btc5.cloud

What this site is, how much data sits behind it, and what to read before you bet.

1. What this is

btc5.cloud is a data & probability reference site for Polymarket's BTC 5-minute Up/Down markets. Using our own oracle and order-book data feeds, we publish direction probabilities, plan signals and live ledgers for every 5-minute window — fully public, read-only, and verifiable day by day.

This is not a copy-trading platform and nothing here is investment advice (see Section 4, Notes).

2. Data: how much we have analyzed

• Market: Polymarket "BTC 5min Up or Down" series — 288 windows a day, one every 5 minutes.
• Settlement basis: the Chainlink 60-second TWAP oracle. Open = the TWAP value at the window's starting second v(b); close = the value at the ending boundary second v(b+300); a tie counts as Up — exactly matching Polymarket's official settlement. All backtests and live determinations on this site use this basis; no exchange price is ever substituted.
• Our own 24/7 collection:
  • Chainlink oracle per-second ticks (same source as settlement);
  • Binance spot per-second ticks (reference);
  • Polymarket CLOB order book: top-of-book bid/ask + 15-level depth, plus every trade.
• Historical scale:
  • 73,107 historical windows (~254 days) in the direction model's backtest;
  • Single-window predictions have full daily records since 2026-07-04 (90+ days) — every day's hit rate is published on the site;
  • Signals such as the 3-leg / 4-leg plans are journaled per window, with backtest-reconstructed segments explicitly marked "incl. backtest".

3. Models & sections

3.1 Single-window plan (direction model)

Features: 21 technical indicators including BOLL(20,2.5), MACD, KDJ, CCI and RSI; a gradient-boosted tree (GBDT) outputs the probability P of the next window closing Up.

Overall backtest accuracy: 51.28% (73,107 windows). Note well: this is only marginally better than a coin flip (see Note 1).

3.2 2-minute plan (t120)

Judges direction from the final 120 seconds of each window only; ~280+ windows per day.

Hit rate over the last 3 days ≈ 79.6% — but by then prices are near their extremes, so high hit rates come with high entry prices and do not imply after-fee profit (see Note 2).

3.3 3-leg / 4-leg plans (martingale chains)

Bets follow the model signal window by window; a chain stops at the first win and doubles after each loss, up to 3 / 4 legs.

Cumulative chain success ≈ 88% (3-leg) / 95% (4-leg). This is the probability of "at least one win inside the chain" — losses concentrate in the tail where the whole chain fails (see Note 3).

3.4 Strong signals (★ / ⚡)

★ strong signal: model probability |P−50%| ≥ 4%. ⚡ band-break reversal: when the latest close exits the BOLL(20,2.5) bands, bet the reversal (56.1% in backtest, ~10 windows/day).

3.5 Live charts & spot

The chart page (/kline) uses our local Chainlink source — the same source as settlement — switchable between 5m/15m, with BOLL/MACD/KDJ indicators.

The spot reference is Binance per-second quotes. "Source age" shows data freshness in real time. The whole site runs on US Eastern Time (ET), aligned with Polymarket.

4. Read before you bet (Notes)

1. Fees decide the line. Polymarket taker fee = 0.07 × shares × p×(1−p), most expensive at 50¢. Buying at 50¢, your long-run hit rate must exceed 51.75% just to break even — below it, the more you play the more you lose. Our single-window model (51.28% backtest) sits below this line; do not treat it as a copy signal.
2. Hit rate ≠ profit. Buying late at 0.8–0.9, even an 80% hit rate can lose money: you win 0.1–0.2 but lose 0.8–0.9, plus fees. Always convert to after-fee EV first — never look at hit rate alone.
3. Chain tails are real losses. 88%/95% chain success means 1 in 8–20 chains fails entirely; one full-chain loss wipes out the profit of several chains. In our history, runs of 4+ consecutive losing legs appear roughly once every 10 days. Before using a martingale chain, know how many doublings your bankroll survives.
4. Settlement is not Binance spot. Chainlink 60s TWAP can differ from spot by tens of dollars short-term; betting the spot direction late can "win the spot, lose the settlement".
5. The market is faster than you. A reversal visible late in the window has usually been priced in 30–60 seconds earlier; a "cheap" price usually implies the same probability. Don't chase late reversals.
6. Liquidity is limited. Top-of-book depth is usually a few hundred to a few thousand shares; large orders slip, and minimum order sizes apply.
7. Data can stall. Our collection occasionally hits upstream hiccups; if the "source age" (top-right, window card) exceeds 10 seconds, stop using the readings.
8. The past is not the future. All backtests and cumulative stats are backward-looking; market structure changes, and every signal has a finite lifespan.

5. Suggested workflow

• First learn the basics of candlestick charts — if you're new, click here.
• Our main model is a direction-probability judgment (it takes no trend/reversal stance), with two explicit reversal signals overlaid — so single-window plans will frequently lose several windows in a row. Never follow plans blindly; treat them as one more indicator, like the chart itself.
• The 3-leg and 4-leg plan pages include trend charts — use them to judge whether current conditions are worth betting at all.
• Use the live spot price to judge whether an open position is worth holding to the end.
• Don't bet big. Start small to learn the flow.

Suggested reference approach:

1. Use the chart's middle BOLL band to judge whether price is trending up or down, and whether it may break through the band;
2. When a plan disagrees with the chart, stay out — or bet only the last 2 legs of each plan to reduce exposure;
3. When losses pile up or you're not in good shape, stop. Never bet angry.

6. Disclaimer

All content on this site is data display and statistical reference only, and constitutes no investment advice or offer. Crypto assets and prediction markets are highly volatile; you may lose your entire principal. Any action taken using this site's information is at your own risk, and any profit or loss is yours alone.