Chapter 30 of 33
The Short Version
Polymarket launched perpetual futures on April 21, 2026 — the same day Kalshi announced its own perps product. Perps are continuous, never-expiring leveraged contracts that track the spot price of an underlying asset. Polymarket's initial roster is BTC, NVDA (Nvidia stock), and gold, with up to 10x leverage, long or short, settled in USDC on Polygon. That's far less leverage than Binance or Bybit offer (up to 125x), but the integration is unique: you can now hedge your prediction-market positions with spot-equivalent derivatives inside the same account. Perps are an experts only product — 10x leverage turns a 10% adverse move into a total wipeout, and Polymarket's implementation is brand new.
Side-by-side: prediction shares resolve $0 or $1; perps track the live spot price forever and cost you funding every 8 hours.
Part 1: What Perps Actually Are
A perpetual future is a derivative contract that tracks the spot price of an asset, has no expiration date, and uses a funding rate mechanism to keep its price tethered to spot. You can go long or short, with leverage, and close whenever you want.
That makes them fundamentally different from prediction-market shares:
| Prediction-market shares | Perpetual futures | |
|---|---|---|
| Resolution | Resolves to $1 or $0 on a binary outcome | Never resolves; you exit when you choose |
| Price range | $0.00 - $1.00 | Tracks the full dollar price of the asset |
| Leverage | None (1x); implicit leverage on cheap shares | Explicit up to 10x |
| Carrying cost | None | Funding payments every ~8h |
| Max loss | 100% of stake | 100% of margin (liquidation) |
| Typical holding time | Days to months | Minutes to weeks |
Launch roster (April 21, 2026): BTC, NVDA (24/7 — the only place to get this), and Gold. Silver, ETH, SOL, and SPY are widely expected next.
Part 2: What's Listed (April 2026)
| Asset class | Symbol | Max leverage | Notes |
|---|---|---|---|
| Cryptocurrency | BTC | Up to 10x | Highest volume; tightest spreads |
| Equities | NVDA | Up to 10x | 24/7 trading on a stock that traditionally closes; novel |
| Commodities | Gold (XAU) | Up to 10x | Classic hedge asset; thin at launch |
Additional assets (ETH, SOL, SPY, oil, silver) are widely expected but not confirmed as of this writing.
Leverage math: $1K margin × 10x = $10K notional → a 10% adverse move = full liquidation. 3x buys ~33% of buffer; 2x buys ~50%.
Part 3: Leverage, Margin & Liquidation
What 10x leverage actually means
With 10x leverage, $1,000 of margin controls $10,000 of notional exposure. Your P&L is calculated on the full notional, not the margin.
- Margin: $1,000
- Notional: $10,000
- BTC price at entry: $100,000 (so you control 0.1 BTC)
- BTC rises 5% to $105,000 → P&L = +$500 → +50% on your margin
- BTC falls 5% to $95,000 → P&L = -$500 → -50% on your margin
- BTC falls 10% to $90,000 → P&L = -$1,000 → liquidation (margin wiped out)
Funding rate
Perps stay near spot because of funding: every 8 hours, one side pays the other based on the perp-to-spot basis.
- Perp trading above spot → longs pay shorts
- Perp trading below spot → shorts pay longs
Typical funding is ±0.01% per 8h (±0.03% per day, ±10% annualised), but in extreme conditions it can spike to ±0.5% per 8h. Funding compounds — a 0.05%/8h funding rate is roughly 5.5% per month, which will eat any weak directional edge.
Liquidation price
Your liquidation price is the spot price at which your margin is fully consumed. Roughly:
- 10x long: liquidates around 10% below entry
- 5x long: liquidates around 20% below entry
- 3x long: liquidates around 33% below entry
- 2x long: liquidates around 50% below entry
Exact numbers depend on maintenance margin, fees, and funding accrued. Polymarket's UI shows your live liquidation price — check it before you open the trade, not after.
Honest comparison: Binance/Bybit still win on depth and track record. Polymarket uniquely wins on 24/7 NVDA and same-account hedging with prediction markets.
Part 4: Polymarket Perps vs Binance / Bybit
| Feature | Polymarket | Binance / Bybit |
|---|---|---|
| Max leverage | 10x | Up to 125x |
| Listed assets | BTC, NVDA, Gold (April 2026) | 100+ crypto pairs; no equities |
| Settlement currency | USDC on Polygon | USDT / USDC |
| 24/7 equities | Yes (unique) | No |
| Regulation | CFTC DCM (US) | Offshore for most users |
| Same-account hedging with prediction markets | Yes | No |
| Track record | Three days old at time of writing | 5+ years |
| Typical fees | TBD (launch; low expected) | Maker 0.02% / Taker 0.05% |
| Liquidation insurance fund depth | New, smaller | Large, battle-tested |
The four strategies that earn their risk: small directional with stops, basis/cash-and-carry, spot hedge, and the native prediction-market delta hedge.
Part 5: Four Concrete Strategies
Strategy 1 — Directional with strict stops
The classic use. Have a directional thesis, size small, use a stop.
- Max 2-3x leverage for swing trades (days to weeks)
- Max 5-10x only for intraday trades with a hard stop
- Stop-loss = entry ± (1× ATR, one day's realised range)
- Position size so that if stop is hit, you lose ≤1% of bankroll
Strategy 2 — Basis trade (cash-and-carry)
When funding is elevated, you can go long spot + short perp (or vice versa) to harvest the funding rate with near-zero directional exposure.
Strategy 3 — Hedging existing spot exposure
You hold 1 BTC in a cold wallet. You're worried about a 2-week drawdown but don't want to sell (tax, lost upside). Short 1 BTC of perp at 2x-3x. Close when the risk passes. This is how institutional crypto desks manage drawdown risk without triggering taxable sales.
Strategy 4 — Hedging prediction-market positions (Polymarket-specific)
This is the native use case. Suppose you have a $5,000 long position in "BTC above $110K by end of May" on Polymarket. Your P&L is part probability, part spot price. If BTC drops 15%, your prediction-market position tanks — not because your probability edge disappeared, but because spot crashed.
Solution: hedge the spot component with a short BTC perp sized to neutralise the delta of the prediction-market payoff. Now your P&L reflects only the information edge in the prediction market, not BTC's directional move.
Sizing matrix. Even on a $100K bankroll, swing trades rarely justify more than 5x. 10x is an intraday-with-hard-stop tool only.
Part 6: Position Sizing for Perps
Everything in Position Sizing still applies — quarter-Kelly, max 5% of bankroll per position, no more than 25% of bankroll across correlated positions. On top of that, perps have two additional constraints:
- Liquidation buffer: your liquidation price should be at least 2× the worst single-day historical move of the asset away from entry. For BTC that's ~20%; so max sensible leverage is around 5x for a swing trade.
- Funding carry budget: estimate 30 days of funding cost. If funding is +0.05%/8h, that's 5.5% monthly. Don't open a long thesis unless you expect the move to exceed funding by a comfortable margin.
| Account size | Sensible single-position margin | Max leverage suggestion |
|---|---|---|
| $1,000 bankroll | $30-50 margin | 2-3x (swing) |
| $10,000 bankroll | $200-400 margin | 3-5x (swing); 10x only for day trades with hard stop |
| $100,000 bankroll | $1,500-3,000 margin | 2-5x; use laddered entries |
Risk catalogue: engine maturity, oracle risk, thin launch liquidity, unknown funding regime, cross vs isolated margin, and Section 1256 tax complexity.
Part 7: Risks Specific to Polymarket Perps
- Engine maturity: the perp engine is days old. Expect bugs, oracle hiccups, and UI issues in the first months. Don't put irreplaceable capital at risk.
- Oracle risk: perps use a price oracle to determine mark price and liquidations. If the oracle misbehaves, you can be liquidated on fake prices. Major exchanges have had these incidents; a newer exchange has more risk, not less.
- Liquidity at launch: early order books will be thinner than Binance. Expect wider spreads and worse slippage on size until market makers arrive.
- Funding regime unknown: the funding-rate algorithm is new. Early weeks may see unusual basis behaviour.
- Cross-margin vs isolated: confirm whether your margin is isolated to a single position or shared across positions. Cross-margin means one bad perp trade can take down your prediction-market collateral too.
- Regulatory unknown for non-US users: perps in some jurisdictions carry different (or outright prohibited) regulatory treatment compared to prediction markets. See Country Guide.
- Tax complexity: perps generally fall under Section 1256 in the US (60/40 treatment), which is different from how prediction-market gains are taxed. Your bookkeeping just got more complicated — see Tax Guide.
Part 8: Early-Access Status (April 24, 2026)
- Rollout is in early access — not every account has perps enabled yet
- Full fee schedule (maker/taker tiers, funding cap, insurance fund) not yet publicly finalised
- No public SDK for perps yet (expected in the py-clob-client roadmap)
- WebSocket feed for perp order books is live but less documented than the prediction-market feed
- Interface is visually similar to prediction markets but with new order-ticket controls for leverage and margin
Part 9: Daily Workflow
- Separate your perps from your prediction-market bankroll mentally — think of them as two accounts
- Open perps only when you have a concrete reason (directional thesis, basis, hedge) — never "just because"
- Always check liquidation price and funding rate before submitting the order
- Size so you can absorb a 2× daily-range adverse move without liquidation
- Set stop-loss limits or time-based exits; don't let a trade drift
- Monitor funding every few hours if carrying overnight
- Close (or reduce) on big known catalysts you didn't model for
- Log every perp trade separately from prediction-market trades for tax
Part 10 — Validated Pro Tips For Polymarket Perps
Habits and rules distilled from basis-trade desks, crypto prop shops, and prediction-market veterans who survived the 2021 wick cycles. None of this is theoretical; each line corresponds to a blow-up someone has already paid for.
- 3x is the ceiling for swing trades. 5-10x is an intraday-only tool with a hard stop. Anything higher is a lottery ticket, not a trade.
- Liquidation buffer ≥ 2× the asset's worst single-day move. For BTC that's ~20% → max sensible leverage is ~5x swing.
- Budget 30 days of funding before opening a carry trade. 0.05%/8h ≈ 5.5%/month. If your thesis doesn't clear that plus fees, it isn't a trade.
- Risk 1% of bankroll to the stop, not 1% to the margin. With 10x, 1% of bankroll is often only $10-20 of margin — that's fine, it's the risk that matters.
- Isolated margin, not cross. A cross-margin blow-up on a BTC perp can seize your prediction-market collateral in the same sweep. Isolate by default.
- Hedge the delta of prediction-market positions, not the notional. A $5K "BTC > $110K by May" position has a spot delta of maybe $1.5-2K — short that amount of BTC perp, not $5K.
- Read funding every 8 hours on carried overnights. Set a phone alarm for funding resets if carrying through weekends.
- Pre-launch engine = paper-size first. Trade $50 margin for two weeks before scaling. Oracle hiccups and UI edge cases only show up in production.
- Track Section 1256 positions separately. US perps get 60/40 mark-to-market treatment — utterly different bookkeeping from Form 8949 prediction-market shares.
- Close on unmodeled catalysts. FOMC, CPI, NFP, major earnings — if you didn't have the event in your thesis, reduce to zero before it prints.
- Never average down into leverage. On a prediction market, averaging down is sometimes rational. On a 5-10x perp, averaging down is how accounts zero.
- Walk away after ±3R on the day. Wicks on leveraged BTC and 24/7 NVDA happen at 3am. Trading tired with margin is the most expensive hobby in finance.
Situation → Action Cheat Sheet
| Situation | Action |
|---|---|
| Funding has spiked to +0.1%/8h (+11%/mo) | Look at basis trade: long spot + short perp to harvest carry; hedged directionally |
| You have a large Polymarket "BTC above $X" long and BTC is weak | Short BTC perp sized to the spot delta of the position; isolate the probability edge |
| NVDA earnings tonight and you're long NVDA perp | Close before the close or trim to 25% — 24/7 NVDA wicks brutally on earnings leaks |
| Engine shows 20 min of lagging oracle prints | Reduce size to zero or move stops far from market; don't try to time an illiquid book |
| Liquidation price within 1.5× of daily ATR | Add margin or reduce size; never hold a liquidation inside one day's range |
| Gold perp 24h volume < $5M | Limit orders only; assume 2-5x normal slippage; size small |
| Funding rate flipped sign mid-carry | Reassess — your free carry just reversed; unwind if the basis thesis broke |
| You're up 100%+ on a 5x intraday swing | Take half off; move stop to entry; let the rest run with zero risk |
What's Next?
- Crypto Trading — the prediction-market side that pairs with BTC perps
- Advanced Strategies — multi-leg structures that combine perps and prediction markets
- Risks — full risk catalogue including perp-specific risks
- Position Sizing — the bankroll math you must use before touching leverage
- Tax Guide — Section 1256 treatment and how to keep perp and prediction-market books clean
