On February 28, 2026, the United States and Israel launched coordinated military strikes against Iran. The Strait of Hormuz, through which 20% of global oil flows, effectively shut down.
Over the next ten days, crude went from $66 to $115. Then on March 9, President Trump declared the war "very complete, pretty much." Oil crashed 30% intraday to $80. Hours later, Defense Secretary Hegseth contradicted him. Prices rebounded toward $93.
Two on-chain platforms processed the volume: Polymarket (binary prediction markets on Polygon) and Ostium (leveraged CL/USD perps on Arbitrum). Same thesis — oil goes up. Very different instruments.
Narrowing to oil-specific markets, the scale gap is significant. Ostium's CL/USD pair processed $191M in notional volume on March 9 and $194M on March 10. Polymarket's CL threshold markets peaked at $8.2M in notional — a 17x gap on oil alone. Over two weeks: $849M in Ostium CL/USD notional vs $49M in Polymarket oil markets.
But that $49M undercounts Polymarket's role in this event. The full Iran-related prediction market ecosystem was far larger: 500+ markets processed $722M in total notional volume — from "Khamenei removed as Supreme Leader" to "Iran closes the Strait of Hormuz" to "US x Iran ceasefire."
These broader geopolitical markets weren't direct oil price bets, but many of them could function as proxies for oil exposure. A "Strait of Hormuz closure" market is fundamentally a bet on oil supply disruption. A "ceasefire by date X" market is a bet on how long the supply shock lasts. Traders could — and likely did — use these as indirect hedges on oil price risk, even though they resolve on geopolitical events rather than commodity prices.
When you include these proxy markets, the picture shifts: Polymarket's $722M in Iran-related notional is comparable in scale to Ostium's $849M in CL/USD. The platforms served different functions — Polymarket offered a rich menu of event-level granularity (regime change, ceasefire timing, supply disruption) while Ostium offered continuous price-level exposure — but the capital deployed was in the same order of magnitude.
Explore Polymarket and prediction markets data
Every price below is an actual on-chain trade execution on Ostium's Arbitrum contracts — not a feed, but what traders actually paid.
The following is not a backtest of a specific trading strategy. It's a structural comparison: given the same macro shock, how did each instrument type behave — and what does that reveal about their strengths and risks? The traders on these platforms included both hedgers (airlines, importers, funds with short oil exposure looking to offset losses) and speculators (directional traders betting on higher prices). The instruments work the same way for both — the distinction is in the trader's intent.
Polymarket had dozens of CL threshold markets: "Will CL hit $X by end of [month]?" Buy the Yes token cheap; if oil crosses the strike before expiry, it pays $1.
In the days before the war (Feb 25–27), "end of February" threshold markets were available at very low prices — the market was pricing almost zero probability of a major oil spike within days. But oil was still at $67 when February ended on the 28th. Every single “end of February” CL market resolved No.
The winning play was the "end of March" markets, created on February 28 — the day strikes began. These were the first available entry for traders reacting to the news, and the extra month of runway turned a total loss into a 4x gain.
Best confirmed return: +300% on the $90 strike — 25¢ to $1 in seven days, no leverage needed. But the identical thesis ("oil will cross $90") with a February expiry would have returned -100%, because oil hadn't moved yet when Feb ended. The instrument's risk isn't the thesis — it's the expiry window.
On Ostium, traders went long CL/USD with leverage and no expiry date. A trader who entered on Feb 25 at ~$66 with 5x leverage and held to the $115 peak on March 9 would have seen roughly +370% on collateral. At a conservative 2x, about +150%. These are illustrative; actual returns varied based on each trader's entry, leverage, and exit.
While prediction markets also let traders reduce positions or buy the opposite outcome as odds shift, the structural difference is in how each behaves during extreme volatility. On Ostium, P&L tracks oil price linearly. On Polymarket, token value reflects a probabilistic assessment — more like an option than a future. A "CL hits $90" token might drop from 95¢ to 70¢ during the $115-to-$80 crash, but if oil finishes March at $89.99, it settles at zero. On Ostium, $89.99 is still $89.99.
Explore Ostium data
The volume comparison understates the most important practical difference: how much a single trader can deploy.
The median Ostium CL/USD order was $1,523 in notional. The median Polymarket oil trade was $30 in notional. Ostium processed 268 orders above $1M, totaling $700M, or 82% of all CL/USD volume. Nineteen exceeded $5M, and the single largest was $13.2M. On Polymarket, the largest oil trade was $500K in notional, and 99% of trades were below $5,758.
But participation tells the opposite story. Polymarket's oil markets attracted 21,390 unique addresses, 22x more than Ostium's 952 CL/USD traders. Across all Iran-related markets, 120,023 unique addresses placed bets. Polymarket processed 125,000 oil trades, driven by a long tail of small, precise bets.
This is the fundamental tradeoff. Ostium concentrates capital: fewer traders, larger positions, deeper liquidity per market. Polymarket distributes participation: orders of magnitude more people expressing views at a smaller scale.
The liquidity fragmentation is partly inherent: each threshold and market is a separate orderbook. A continuous perp concentrates all directional flow into one instrument. But the current depth gap also reflects where prediction markets are in their adoption curve, rather than a permanent ceiling. As infrastructure matures, depth should improve, even if structural fragmentation remains.
This is where Ostium reveals something more: how the entire platform was positioned across the macro landscape.
By March 10: CL/USD was 98% long. Gold was 98% long. S&P 500 was 94% short. The textbook stagflation trade.
The S&P went from 98% long to 25% long in under a week — a complete reversal. None of this exists in prediction market data. Each Polymarket market is isolated; you can't see how the same capital is positioned across oil, gold, equities, and crypto simultaneously.
Both instruments let traders express their conviction about oil price, whether as a hedge against real-world exposure or a directional bet on the spike. But they do so in structurally different ways, and those differences determine who can use them, at what scale, and with what risk.
Polymarket's strength is granularity. No other instrument lets you bet precisely on "will the Strait of Hormuz close by March 31" or "will there be a ceasefire by March 15." The $722M in Iran-related notional volume across 500+ markets shows genuine demand for event-level exposure. For small, precise, high-conviction bets, the capital efficiency is unmatched: a 25-cent bet returning $1 is a 300% return with zero leverage and no liquidation risk. Polymarket's weakness is liquidity fragmentation and the expiry constraint. At least for now, you can't move large sizes without considerable slippage, and binary resolution with a fixed deadline means timing matters as much as direction.
Ostium's strength is liquidity and flexibility. A single CL/USD market absorbed $191M in daily volume with sub-basis-point price impact for large trades. No expiry means no timing trap. Cross-asset exposure means you can express a full macro view — long oil, short equities — rather than isolated binary bets. And real-time positioning data provides macro context invisible to prediction markets. Ostium's weakness is that leverage cuts both ways. This is where the big, concentrated bets happen and the risk per trader is far higher than Polymarket's retail-sized bets.
The Iran oil-shock episode highlights how prediction markets and perps serve different parts of the capital spectrum. Institutions and large traders who need to move serious capital gravitate toward perps: deep liquidity, continuous exposure, and positions in the millions. Retail participants and those seeking precise, event-level bets gravitate toward prediction markets: lower capital requirements, no leverage risk, and a rich menu of granular outcomes.
They can also work in combination: prediction markets deliver convex upside, while perps preserve directional exposure when timing is off. But in practice, the trade is still fragmented across platforms and chains, with separate collateral, no shared margin, and no portfolio netting.
This is starting to change. Hyperliquid's HIP-4 — now live on testnet — puts outcome contracts and perpetual futures on the same margin engine, enabling shared collateral and portfolio-level risk management across both instrument types. If this kind of infrastructure had existed during the Iran crisis, a trader could have held a "Hormuz closure" outcome and a CL/USD perp long in the same margin account — one chain, one risk engine, one pool of capital.
For teams looking to go deeper, Dune provides access to the underlying datasets behind both prediction markets and perpetual DEXs — from market-level event activity to perp positioning and volume — available directly in Dune, via API, or integrated into internal workflows.
All data in this analysis is sourced from on-chain records indexed by Dune.
Looking to use Dune for your company?
