Futures contract selection beyond ES/NQ¶

Date: 2026-10-05

Scope and purpose¶

Read-only research to inform a buy/build decision on data sourcing and strategy research priorities. Question: beyond ES and NQ, which CME-family futures contracts are good candidates for systematic (intraday/swing) strategy research — enough movement for an edge to clear transaction costs, but still liquid enough to trade and to source decent historical data for.

Methodology note: this is desk research synthesized from CME Group's own press releases/market-data pages (authoritative for volume), and from trading-education and prop-firm blog sources (useful for range/ATR figures and style framing, but not independently verified — flagged below). No qkt backtests were run for this doc; it is a market-selection literature review, not an edge claim.

1. Volatility / liquidity comparison¶

Volume — CME-reported, authoritative¶

CME Group's own press releases give real, dated ADV (average daily volume) figures. From the September/Q3 2026 record release and related monthly releases:

Overall CME ADV: 31.8M contracts/day in September 2026 (+22% YoY); Q3 2026 ADV 29.4M. CME Group press release, Sept/Q3 2026
Interest rates: Interest Rate ADV +22% to 16.2M contracts/day; U.S. Treasury futures+options ADV +42% to 9.5M/day. ZN (10-Year T-Note) specifically: ADV ~2.5–2.9M contracts/day across different 2026 reporting periods (March +21% to 2.9M; Q3 +27% to 2.5M; September +51% to 2.6M) — ZN is one of the most liquid futures contracts on earth. CME press releases, 2026
Energy: Energy ADV +37% to 3.1M/day; Henry Hub Natural Gas (NG) ADV +20% to 602K/day; Micro WTI Crude Oil ADV +376% to 240K/day (micro growth story, see §3). CL (crude oil) full-size ADV is commonly cited around ~700–800K contracts/day (blog-sourced, consistent with CME's energy-complex totals but not a direct CME per-symbol citation — treat as approximate).
Agriculture (context, not in this task's scope): corn ADV +62% to 461K/day, soybeans +35% to 289K/day — included only to show agriculture is a real volume category too, if ever considered.
FX: CME FX futures+options hit 3.3M contracts on a single September 2026 day (3.2M of which futures) — fourth-highest ADV since 2014. 6E (Euro FX) is the dominant FX future; third-party (non-CME) sources cite 6E trading 200K+ contracts/day in active sessions — unconfirmed against a primary CME per-symbol figure, but directionally consistent with 6E being the most liquid currency future by a wide margin over 6J (yen) and 6B (pound).
Equities: ES ADV cited at ~1.6M contracts/day (blog-sourced aggregation, not a direct CME quote pulled in this pass, but plausible given ES's known dominance). MNQ now regularly outtrades full-size NQ per 2026 reporting — a genuinely interesting liquidity-migration fact (see §3).
Metals: Gold (GC) ADV cited around 300K+ contracts/day.
RTY, YM, HG (copper), SI (silver), 6J, 6B, ZB — no reliable per-symbol ADV figure was found in this pass from a primary CME release; these are meaningfully less liquid than ES/NQ/CL/GC/ZN/6E but are still exchange-top-30 products traded by CTAs and prop firms daily. Treat any specific number for these as unconfirmed until pulled directly from CME's exchange-volume or monthly volume report PDFs.

Actionable next step if precision matters: CME publishes a monthly volume report PDF (Web_Volume_Report_CMEG.pdf) and live exchange-volume / metals-volume / equity-volume pages with exact per-symbol ADV and open interest — worth a direct pull before finalizing a data-vendor spend decision, since blog aggregations in this report were inconsistent on CL/GC specifics.

Daily range / ATR — blog-sourced, unconfirmed, directional only¶

No primary-source (CME) ATR/range dataset was found; the figures below are from trading-education sites (atlasalgo.io, bulltraders.com, volatilitybox.com, tradingsim.com) and should be treated as rough, time-varying approximations, not research-grade constants:

Contract	Typical daily range (cited)	Character (cited)
ES	~40–80 index pts	steady, trendable but lower raw-point range than NQ
NQ	~200–400 index pts (NQ moves ~1.3–1.5x ES on a % basis)	highest vol of the index complex, favored for trend/breakout
RTY	~20–35 pts, but highest % volatility among index futures (small-cap beta)	choppier, gap-prone around macro releases
CL	~$1–3/barrel ($1,000–$3,000/contract)	large absolute moves around EIA inventory, OPEC headlines; historically trend-until-it-doesn't, dangerous for naive mean reversion on outright price
GC	~$20–50/oz ($2,000–$5,000/contract)	macro/inflation/dollar driven, distinct catalyst set from equities

These numbers are directionally useful (NQ > ES in raw-point and %-range terms; RTY has outsized %-volatility despite modest point range; CL and GC both move enough per day to clear realistic commission+slippage on a well-sized intraday/swing system) but should be re-derived from qkt's own backtest data (ATR indicator over real history) before being used to size strategies, per qkt/docs/reference/dsl/*.md indicator conventions.

2. Style classification — trend vs. mean-reversion vs. avoid¶

Trend-following — well-documented, cross-sourced: - The managed-futures/CTA industry's live track record is built almost entirely on time-series momentum/trend-following across commodities, currencies, and rates (not single-stock-index trend). CASAM CISDM CTA index compounded ~14.5%/yr 1980–2010 vs ~7% for US stocks; Barclay CTA Index +~14% in 2008 while S&P 500 fell ~37%; SG Trend Index +~27% in 2022 while stocks and bonds both fell. Candriam, ETF.com/Swedroe, CME CTA comparison paper — this is the strongest, best-cited evidence point in this report: currencies and commodities (not equity indices) are the historically-documented home of systematic trend-following edge. - NQ is repeatedly cited in trading-education material as the equity-index contract most "favored for trend following" due to its higher volatility — but this is a style preference claim from blogs, not a CTA-index-level statistical claim like the one above.

Mean-reversion — mixed, market-structure dependent: - Crude oil outright price mean-reverts unreliably ("trends until it doesn't" — naive dip-buying eventually hits 2014 or 2020 and doesn't recover). Reliable mean reversion in oil lives in spreads (WTI–Brent, calendar spreads vs. inventory), not in CL outright — relevant if considering CL for a mean-reversion system rather than trend/breakout. Substack: Do Commodity Futures Trend or Mean Revert? - Academic literature (ScienceDirect) on energy-futures mean reversion reports Sharpe ratios >2 for WTI and Natural Gas in some specifications — a peer-reviewed-adjacent source, but methodology/period not verified here; treat as a lead for further reading, not a settled fact. ScienceDirect: Trading on mean-reversion in energy futures markets - Major FX pairs are commonly cited as the "most stationary" (mean-reverting) instruments among liquid futures — consistent with 6E/6J/6B being useful for range-bound/mean-reversion systems in calm regimes, though they also trend hard during macro (rate-divergence, carry-unwind) regimes.

Illiquid/avoid for algo execution: - Lumber and orange juice (FCOJ) are the two most commonly named "stay away" thin markets in retail/prop-firm education: low open interest, bids/offers that evaporate during moves, large slippage from small orders. Neither was in this task's candidate list, but worth noting as the negative baseline — none of ES/NQ/RTY/YM/CL/NG/GC/SI/HG/6E/6J/6B/ZN/ZB/VX fall into this "avoid" tier; they are all CME-top-tier or CFE-flagship products. - VX/VIX futures do not trade on CME — they trade on Cboe Futures Exchange (CFE), a separate exchange from the CME-family products in this task. Flagging explicitly since the task scoped "CME-family." VIX futures are liquid (tens of thousands of contracts/day) and have a Mini-VIX (1/10 size) contract, but margin is high (~$4,000–$6,000/contract cited) and the term structure/contango dynamics make backtesting materially harder than a linear product — a reasonable "stretch" addition but outside strict CME scope and structurally different (cash-settled on a forward-implied-vol index, not a physical/financial underlying). Optimus Futures VIX guide, Macroption VIX futures

3. Micro contracts — confirmed specs¶

Confirmed from CME Group's own contract-spec pages:

MES (Micro E-mini S&P 500): $5 × S&P 500 index, tick 0.25 pts = $1.25/tick. CME MES
MNQ (Micro E-mini Nasdaq-100): $2 × Nasdaq-100 index, tick 0.25 pts = $0.50/tick. CME MNQ
Both are exactly 1/10th the notional of their full-size E-mini (ES/NQ) counterparts, same trading hours (Sun–Fri 5pm–4pm CT, 1hr daily halt).

Blog-sourced (not independently re-verified against CME spec pages in this pass, but internally consistent and plausible) for the rest:

M2K (Micro Russell 2000): $5 multiplier (1/10 of RTY).
MYM (Micro Dow): $0.50 multiplier (1/10 of YM) — smallest $/tick of the micro-index suite.
MGC (Micro Gold): $10 multiplier, tick 0.10 = $1.00/tick (1/10 of GC).
SIL (Micro Silver): $1,000 multiplier, tick 0.005 = $5.00/tick.
MCL (Micro WTI Crude): $100 multiplier, tick 0.01 = $1.00/tick (1/10 of CL).

Notable liquidity-migration fact (from CME's own 2026 reporting, cited above): MNQ now regularly out-trades full-size NQ, and Micro WTI (MCL) ADV was up 376% YoY to 240K contracts/day in September 2026 — micros are not a toy product anymore; for several contracts they are now where the retail/ small-systematic flow actually is. This directly supports testing at micro size before scaling, both for realistic position sizing and because the micro order book itself is now deep enough to be representative.

Recommendation for this task's framing: confirm exact current specs (tick size, margin, multiplier) directly against CME's contract-spec pages per symbol before wiring into qkt's instruments.yaml, since multipliers and tick sizes are exchange-set and this doc's secondary sources should not be the system of record for cost modeling.

4. Data/connectivity coverage¶

Databento's GLBX.MDP3 dataset (CME Globex MDP 3.0) is the single feed for essentially everything in this comparison: it covers all CME, CBOT, NYMEX, and COMEX-listed futures, options, spreads and combinations in one feed — 650,000+ symbols, tick-level with full order-book depth. CME explicitly lists ES, CL, 6B, ZN among its own example instruments on that feed. Databento GLBX.MDP3

Practical implication for the buy/build decision: picking a vendor for ES (e.g., Databento GLBX.MDP3) is picking a vendor for essentially the entire candidate list — NQ, RTY, YM, CL, NG, GC, SI, HG, 6E, 6J, 6B, ZN, ZB and all their micro equivalents (MES, MNQ, M2K, MYM, MGC, SIL, MCL, etc.) are the same exchange group (CME/CBOT/NYMEX/COMEX) and therefore the same feed/dataset — there is no per-contract vendor-selection problem within CME-family products. The one named exception is VX/VIX, which is CFE (Cboe), not CME — if VIX futures are wanted, that is a separate data source from GLBX.MDP3 (Cboe's own feed or a vendor with CFE coverage), not a line item within the CME dataset purchase.

This was not independently cross-checked against a second vendor (e.g. CQG, Barchart, IQFeed, Polygon) in this pass — Databento was the only vendor researched because it was named in the task. If a second data-vendor quote is wanted, same question (coverage breadth across the symbol list) should be checked explicitly, but the structural point — one CME membership/feed = coverage of essentially the whole non-VIX candidate list — should hold across vendors since it reflects exchange structure, not vendor-specific packaging.

5. Recommendation¶

For a solo systematic trader already running ES + NQ wanting 1–2 more contracts for diversification, the most consistently recommended pair across prop-firm education and trading-literature sources is:

CL (Crude Oil) and GC (Gold) — or their micros (MCL, MGC) at small-account size.

Why this combination, synthesizing the above: 1. Different catalyst set, not just different beta. ES/NQ move on equity- risk and rate-expectations flow; CL moves on EIA inventories, OPEC+ supply headlines, and geopolitical risk; GC moves on real yields, dollar strength, and safe-haven flow. When equity indices chop sideways, CL or GC frequently still have a clean directional day — genuine diversification of opportunity, not just of correlation statistics. 2. Enough raw movement to clear costs. Both show daily ranges in the thousands of dollars per full-size contract (cited ranges above), well above typical commission+slippage for either full-size or micro contracts. 3. Liquid enough to trade and to get clean data for. Both are top-tier CME products (energy and metals volume leaders respectively), both have mature micro equivalents (MCL, MGC) for realistic small-size testing, and both are covered by the same GLBX.MDP3 feed already needed for ES/NQ — zero incremental data-vendor cost to add them. 4. Historically documented edge type is different from equities. The CTA/ managed-futures literature's trend-following track record is built on commodities and currencies, not equity indices — so adding CL/GC is also adding exposure to the asset classes where systematic trend-following has the longest, best-documented live track record, as opposed to adding a third/fourth equity-index future that is highly correlated with ES/NQ (RTY and YM move with ES/NQ most of the time; marginal diversification value is lower).

Runner-up candidates, with caveats: - 6E (Euro FX) — most liquid currency future, different driver (FX/rate differential), commonly cited as mean-reversion-friendly in calm regimes; reasonable 3rd addition if currency exposure specifically is wanted, but the task's own sources suggest CL/GC is the more commonly cited "first two beyond ES/NQ" pairing. - ZN (10-Year Note) — extremely liquid (one of the most liquid futures on earth per 2026 CME data), but behaves more like a slow-moving rates/macro instrument; good for rates-view strategies, less obviously suited to intraday breakout systems than CL/GC. - NG (Natural Gas) — high realized volatility and decent liquidity (ADV 602K/day per CME September 2026 data), but literature consistently flags it as prone to violent, news-driven (storage report, weather) gap moves that are harder to risk-manage systematically than CL; a plausible 3rd pick for a trader specifically wanting more volatility, with the explicit caveat that gap risk is higher. - VX/VIX — interesting but out of CME-family scope (Cboe/CFE), different data vendor, and structurally harder to backtest (term structure, contango roll) — not recommended as a near-term addition given the stated goal of adding "1–2 more contracts" cleanly.

Open questions / unconfirmed items to close out before committing spend¶

Exact current per-symbol CME ADV for RTY, YM, HG, SI, 6J, 6B, ZB — pull directly from CME's monthly volume report PDF or exchange-volume page rather than relying on this doc's blog-sourced approximations.
Independent ATR/daily-range dataset (ideally computed from qkt's own ingested history via the atr/range indicators in qkt/docs/reference/dsl/indicators.md) rather than trading-blog figures, before using range numbers for position sizing or strategy-fit decisions.
Second data-vendor coverage check (CQG/Barchart/IQFeed/Polygon) if Databento pricing or terms end up being a blocker — the "one feed covers everything" conclusion in §4 should hold structurally but was only checked against Databento in this pass.