RELEASE BRIEF · 0.2.0 → 0.2.2

AQ Studio 0.2

Research, execution and performance

2 November 2026 · For developers and quant researchers

The 0.2 series connects reusable Rust research, detailed execution analysis and live strategy monitoring. Version 0.2.2 adds more dependable strategy controls, smoother reading and navigation, and measured improvements to engine and chart processing. This brief consolidates the user-facing changes across all three releases.

Cloud account transition
Previous cloud accounts, memberships, papers, strategies and history were not imported into the new cloud service. Register again to use cloud features. Local projects and saved local backtests are not deleted by this transition.

01 · Research that stays with your project

Create multiple notebooks per strategy project, keeping notes, Rust code, saved results and historical data together.

• Run and iterate: Execute a cell or the whole notebook; inspect status, elapsed time and inline errors. Reorder cells, retain unsaved drafts across tabs, and resize or collapse the session panel.

• Reuse strategy code: Load research tools prepares project imports and backtest helpers. Reuse alpha models, indicators and custom Rust components; development compilation and saved caches reduce repeat setup.

• Explore visually: Tables, candles, line overlays, separate indicator panels, signal markers, scatter plots, distributions and heatmaps; compare symbols from a common starting point.

• Compare experiments: Open saved backtests to compare returns, drawdowns and insights, or run an independent research backtest on saved historical data without live orders.

• Choose data precisely: Select a named feed, symbols, timeframe and UTC range. Historical CSV feeds support symbol, timeframe, timestamp and OHLCV columns; missing timeframes are not silently substituted.

• Manage the workspace: Search, group and duplicate feed/broker connections. Research selectors exclude execution-only brokers. Deletion and cache-clearing previews, progress notifications and new research guides support longer sessions.

Availability: notebooks run Rust locally. Python and remote notebook sessions are not supported. Initial setup may take several minutes; CSV feeds are historical-only.

02 · Analysis with evidence behind the numbers

Attribution

Alpha contribution by strategy type and symbol, contributing insight families, and drill-downs from summary charts to individual trades.

Accounting and risk

Accounting reconciliation, gross-to-net costs, turnover, exposure, drawdown episodes, correlations and tail-risk contribution. Frozen-fill cost scenarios compare costs without claiming a new execution simulation.

Execution quality

Observed fills, inventory, costs and marked performance; qualified maximum adverse/favourable excursion and exit efficiency where price-path evidence exists.

Result exploration

Sortable backtest and insight tables, newest-first results, pinned columns, clearer sweep IDs and deletion confirmations. Read-only SQL-style filters with autocomplete cover local and cloud metadata; invalid queries retain valid results and pause bulk sweep actions.

Readable results

Grouped metrics, sticky controls, expandable methodology and retained content during refresh. Version 0.2.2 wraps long descriptions and distinguishes loading, sign-in, failed requests and missing evidence.

Live risk: automatic estimates use one-minute returns over a rolling two-hour window and require 20 qualified minute returns. Gaps are excluded; separate daily research needs longer history. Account-level metrics remain distinct from strategy attribution, and annualised ratios stay unavailable when the recorded calendar cannot support them.

Interpretation: missing prices, history or valuation inputs are explained rather than shown as zero. Incomplete paths are identified. Full decision-funnel, walk-forward and strategy-removal comparisons are not included; market replay replays recorded events, not counterfactual execution.

03 · Clearer live trading and chart monitoring

• Charts and equity: Entry, protection and closed-trade annotations update with insight state. Dotted lines connect closed entries/exits. Equity history has improved ordering, baselines and 24-hour/all-time selection; point limits are configurable.

• Monitoring: More consistent prices, P&L, cash and allocation updates. Stopped strategies remain visible after running strategies. Parent/child labels, parent-ID navigation, split-order history, console scrolling and chart controls are clearer.

• MT5 and order handling: Improved order-event matching, cancellation reconciliation, TP/SL commands, nested insights and broker quantity limits, including split orders. Account baselines start from the first broker observation and remain stable across restarts of the same strategy/account.

• Continuity: Live updates recover more reliably from interruptions and stale or duplicate information. Temporary market-data delays no longer immediately stop a running strategy. Local-first backtest loading handles expired cloud sessions; strategy renaming preserves compiled-program lookup.

• 0.2.2 controls: A timed-out Stop continues tracking the relevant run to its final state. Older updates cannot revive a stopped run. Startup feedback distinguishes early exit from timeout, and Insight Flow accepts integer and fractional volume totals consistently.

Eligible inactive broker-account reservations now offer recovery after you confirm the old process has stopped. Recovery does not stop processes or change broker orders or positions. Only one strategy process may use a broker account at a time.

04 · Share research through AQ Insights

Turn a notebook into a paper draft, or attach selected cells to an existing draft. Choose the code and saved output samples to include; readers do not need to execute the notebook.

• Readable research: each notebook cell keeps code and outputs together in a movable block, with cell-level comments and room for surrounding explanation.

• Interactive evidence: notebook and backtest charts support zoom, pan, indicator panels and signal markers in AQS and on the website. Backtest previews show up to 500 equity points, with drawdown, trade and insight counts when available; saved analysis travels with shared backtests.

• Steadier reading: author and engagement controls remain accessible while scrolling. Version 0.2.1 improves recovery from temporary loading failures; 0.2.2 updates reactions without jumping the article and prevents repeated clicks while saving.

• Safer recovery: confirmed backtest-download transaction failures can retry without a second charge; ambiguous responses are not blindly replayed.

Sharing control: notebook inserts use saved snapshots. Inserting alone does not upload them; review selected code, outputs and data permissions before saving or publishing. Failed or stale results are not represented as current research.

05 · A more consistent daily workspace

• Quick navigation: Cmd+K / Ctrl+K finds projects, notebooks, strategies, servers, tabs and settings. Use @username for profiles, $BTCUSD for charts and ?query for AQ Insights. General workspace searches stay local; notebook code and document contents are not indexed.

• Layout and feedback: Refreshed nodes, properties, file trees and hierarchy controls; collapsible pipelines and automatic layout. Drag tabs to reorder, close other/right-hand tabs, and track collapsible research notifications. In 0.2.2, selecting a tab reveals it without vertical page movement.

• Accounts and membership: Cleaner settings, profile, privacy and update screens. Cloud session and membership details sit in Settings; Hobby includes one saved cloud strategy. Quota prompts and website billing links make limits easier to manage. Cloud deletion and storage credits improve in 0.2.1.

• Names and validation: Visible counters and clear errors explain input limits. Over-limit saves are rejected; existing content is not renamed or truncated. Limits include 20 characters for usernames/display names, 50 for first/last names, 100 for bios, 254 for email, three interests of 25, and 64 for project/strategy/server/notebook/connection names.

• Sign-in and protection: Improved session restoration after sleep or connection loss. New usernames use lowercase. Saved credentials and SSH keys use encrypted, OS-protected storage; automatic launch logs redact secrets.

06 · Performance, with a defined comparison

Engine study · Apple M5 Max · 2 November 2026. Five alternating before/after pairs used matching workloads, compiler, dependencies and release settings. Each run used three warm-ups and 15 samples; reported engine values select the slowest median across five runs.

701,213 orders/s in the direct PaperBroker test. ¹ Gains are against the paired pre-optimisation candidate. A like-for-like comparison with September 0.2.0 has not been established.

Complete lifecycle checks

Lifecycle checks preserved identical economic outcomes. The offline live median was effectively unchanged and includes a five-second wait; it is not broker execution latency. Engine tests are synthetic and local, excluding end-to-end MT5, network and UI latency. Peak suite RSS was 120.8 MiB, 3.5% above baseline; no peak-memory improvement is claimed.

Analysis and chart processing

Paired component tests used identical inputs and checked outputs. Backtest/research values use debug builds; timeline values measure native processing. Overlay values exclude canvas painting/layout; research values exclude compilation/rendering. These are component results, not an application-wide speedup. Larger histories use additional setup time and retained memory to accelerate repeated chart updates.

Validation: the optimisation snapshot passed 2,020 correctness tests, browser chart checks and a release WebAssembly build. Dated measurements identify tested source snapshots; they are not measurements of every distributed installer.

07 · Availability and upgrade essentials

Version 0.2.2 is available through the normal download/updater channels, with matching remote runtime packages.

1. Stop and align versions. Stop strategies gracefully; update AQ Studio, matching remote runtime packages and the MT5 bridge EA. Rebuild strategies against AQE 0.2.2 and regenerate old exported deployment commands.

2. Validate custom code. Broker and insight interfaces have changed since older versions. Rebuild and test custom components with a paper/demo account before live use. AQE remains independently buildable and source-available; consult its licence before reuse.

3. Preserve older research and run data. Older development journals and backtest formats are not automatically converted. Retain their matching binaries/storage until pending work is settled; unsupported files are rejected, not reset. Do not delete local databases as an upgrade step.

4. Restore cloud access. Register again for the new cloud service as noted above. Keep the OS credential store available and unlocked, back up SSH keys, and keep exported launch configuration private.

5. Keep research reproducible. Run only trusted notebooks: code has your local account’s permissions. Retain original data/results, recheck saved outputs after code or data changes, and confirm market-data sharing rights.

Release basis: consolidated from the 0.2.0, 0.2.1 and 0.2.2 release notes, with the dated optimisation and validation results. This remains an early release.

https://algoquantstudio.com/

Happy Trading!