← Back to Blog · 2026-09-17 · 8 min read · Strategy Case Study

You've probably run this test a dozen times on ES or crude oil: a 20/60 EMA crossover, daily bars, ten years of data. Sometimes it works, mostly it chops. And at some point you've wondered — what would the same simple logic do on a market most Western traders have never touched? Chinese commodity futures are among the most actively traded contracts on earth, and PTA is one of the workhorses of the Zhengzhou Commodity Exchange. So we did what any curious trader would do: we pulled the history, defined clean rules, and let the data talk.

Here's what a moving average crossover looks like on PTA — the good, the ugly, and the very China-specific traps that don't show up in your usual backtest.

Why PTA, and What You're Actually Trading

PTA (purified terephthalic acid) is the intermediate chemical that feeds polyester — clothing, bottles, packaging. Its price chain runs roughly from crude oil to naphtha to paraxylene (PX) to PTA to polyester. That matters for a trend trader: PTA inherits crude's big directional moves, but it also has its own supply dynamics from Chinese plant turnarounds and capacity cycles, which means it doesn't just mirror WTI. It trends when crude trends, and it mean-reverts when its own supply story dominates.

For retail traders looking at China futures, PTA has real practical appeal:

• It's cheap to trade. At 5 tons per lot and prices historically ranging from the low-3,000s to around 7,000+ yuan per tonne, notional value per lot is roughly 15,000–35,000 yuan — a few thousand US dollars. This is one of the most accessible Chinese commodity futures for a small account.
• It's liquid. The main contracts regularly trade hundreds of thousands of lots a day, so slippage assumptions don't need to be heroic.
• It has a genuine macro story. Crude shocks, the 2020 oil collapse, the 2021–2022 energy squeeze — PTA was dragged through all of it.

The contract, in numbers

One detail worth internalizing: at 10 CNY per tick, a 100-point move — routine during a crude-driven week — is 1,000 CNY per lot against a notional of maybe 25,000 CNY. PTA is a small contract with a big lever. Size accordingly.

Setting Up the Test: Rules That Leave No Room for Cheating

Every failed backtest I've seen dies from ambiguity, so let's pin everything down.

• Signal: Go long when the 20-day EMA crosses above the 60-day EMA; go short on the opposite cross. Always in the market — the classic version.
• Data: Daily bars from a back-adjusted continuous contract, stitched using the actual roll dates when the 1/5/9 main contract changes. If you backtest on a naive concatenation of contracts, the artificial price jumps at roll points will invent trades that never existed.
• Entry/exit: Enter on the next day's open after the signal prints on the close. No intrabar peeking.
• Costs: ZCE fees on PTA are low — a few yuan per lot — but commissions aren't the real cost. We assumed 1 tick of slippage per side plus roughly 2 ticks of roll cost per roll, which is conservative but honest for a retail-size order.
• Risk: Fixed fractional sizing at roughly 1% account risk per trade based on a 2×ATR(20) stop, with the stop trailed. The raw crossover without a stop is a different (and worse) animal — more on that below.

Test window: roughly a decade of history, which conveniently includes the 2020 crude collapse, the 2021 Chinese commodity rally, and the 2022 energy volatility. If a trend system can't prove itself across that stretch, it isn't a trend system.

What the Backtest Actually Showed

Let's be straight: this is not a strategy case study where the equity curve goes up and to the right and everyone rides off into the sunset. The honest picture looks like this.

The long flat stretches hurt more than the losses

PTA spends long periods range-bound in a band of a few hundred yuan per tonne while Chinese capacity additions cap any sustained rally. During those phases, the 20/60 crossover does what crossovers always do in ranges: it buys strength and sells weakness, repeatedly, and bleeds. Counting signals over a decade, the majority of crossovers whipsawed out within days. That's not a PTA flaw — it's the universal tax on trend-following — but PTA's capacity-driven ranges collect that tax aggressively.

The winners came from crude, not from PTA itself

The periods where the system made real money line up almost perfectly with macro energy moves: the crash and rebound around the 2020 oil collapse, and the inflation-driven commodity surge through 2021–2022 — the same wave that took Chinese thermal coal futures up so violently that the exchange repeatedly raised margins and trading limits to cool it down. When the whole petrochemical chain is repricing, PTA trends beautifully, and a daily crossover catches a meaningful chunk of it. When the story is domestic PTA capacity, the system churns.

Practical takeaway: a PTA crossover is secretly a crude-oil-volatility strategy wearing a chemical-industry costume. If you wouldn't trade trend-following on crude, don't expect magic on its downstream cousin.

The stop matters more than the signal

Running the raw always-in crossover versus the ATR-stopped version produced visibly different risk profiles. Without a stop, single trades could sit through multi-hundred-yuan adverse moves — survivable on paper, uncomfortable at 10 CNY a tick with real money. The 2×ATR stop cut the worst drawdown roughly in half at the cost of a few whipsaw exits. For a retail trader, that trade is worth making every time.

The China-Specific Traps Your Backtest Software Won't Warn You About

This is the section that matters if you're coming from Western markets. Chinese commodity futures punish imported assumptions.

Holiday gaps are a different species

Chinese New Year and the October Golden Week close the exchange for about a week. Global markets keep moving. A crossover signal that fires the day before the holiday can open into a gap of several percent — and daily price limits can lock the market so you can't exit at all. Any backtest on Chinese commodity futures must model the gap, not average it away. We flagged positions held into a major holiday and reduced size; the backtest results improved simply by surviving the tail better.

The exchange intervenes, and intervention changes your strategy's math

During the 2021 thermal coal episode, ZCE and other Chinese exchanges responded to runaway rallies by hiking margins, widening price limits, and restricting certain accounts. PTA itself wasn't the epicenter, but the same toolkit gets applied across the board during speculative surges. For a crossover trader this means: margin can jump mid-trend, forcing position cuts your backtest never anticipated. Keep your leverage assumptions conservative — if your sizing only works at minimum margin, it doesn't work.

The night session changes what "daily" means

ZCE PTA trades a night session until 23:00 Beijing time, and the daily bar convention folds that session into the trading day. If your data provider builds daily bars differently than the exchange convention, your crossover signals can shift by a day — which is enough to wreck a backtest's reproducibility. Verify your bar construction against the exchange's official settlement data before trusting any result.

Contract roll is not optional housekeeping

Because liquidity concentrates in the 1/5/9 months, the main contract jumps three times a year with gaps between expiring and new deliveries. Roll on a fixed rule (for example, when open interest in the next main contract overtakes the current one), apply back-adjustment for the signal series, and account for the spread cost. Skipping this is the single most common way Western traders get fantasy results on China futures.

Making the Strategy Survive: Practical Adjustments

If you want to trade something like this for real, here's what the historical evidence supports:

• Add a trend-strength filter. Only take crossovers when the 60-day EMA slope exceeds a threshold, or when ADX is above the low-20s. This kills a large share of the range-bound whipsaws at the cost of entering trends a little later. On a market that ranges as much as PTA does, that's a good trade.
• Watch crude as a regime filter. PTA's best trending stretches coincided with crude's. Some traders only enable long crossovers when crude's own daily trend is up, and vice versa. It's a blunt filter, but the logic is sound — you're trading the chain, not the molecule.
• Size for the gap, not the ATR. If a normal stop is 2×ATR but a holiday gap can be 5%+ against you, your true risk is the gap. Cut size until the worst realistic gap is a tolerable loss.
• Trade the main contract only. The back-months look tempting on paper; their spreads and thin books will eat the edge alive in execution.
• Expect the fee structure to be your friend. ZCE transaction costs on PTA are genuinely low, so unlike high-frequency strategies, a daily crossover isn't strangled by costs — slippage discipline is what protects you, not the commission schedule.

How to Test This Yourself (Properly)

If this sparked something, don't take our word for it — run it. The workflow we'd recommend:

• Source daily data aligned to ZCE's official settlement convention, back-adjusted across the 1/5/9 rolls.
• Write down every rule — signal, entry timing, stop, roll rule, holiday handling — before you look at equity curves. Ambiguity is where self-deception lives.
• Include costs of at least 1 tick slippage per side and realistic roll costs.
• Stress the tail: manually check what the system was holding into Chinese New Year and Golden Week each year.
• Forward-test on a demo or small live size before scaling — Chinese market microstructure has a way of humbling backtests.

And if you want to find out whether your system holds up when it's your own capital on the line — with real data, real contract specs, and a defined risk framework — that's exactly what we built XS Select for. It's a futures evaluation platform for China's markets, starting from $29, where you can take a strategy like this PTA crossover from spreadsheet to a structured trading challenge. No promises of easy profits — just an honest proving ground for a market most traders have never tested.

The Bottom Line

A moving average crossover on PTA futures is not a money machine, and any analysis that told you otherwise would be selling you something. What the history does show is a small, liquid, crude-linked contract that trends hard during global energy dislocations and chops mercilessly otherwise — with a set of China-specific mechanics (holiday gaps, exchange intervention, contract rolls, night sessions) that will break a lazy backtest before it breaks your account. Respect those mechanics, filter for regime, size for the gap, and you'll have a strategy that's boring most of the time and interesting exactly when it should be. In this business, that's about as good as an honest edge gets.