ibkr-trading

name

ibkr-trading

description

Comprehensive Interactive Brokers algotrading knowledge base covering TWS API 10.45 architecture, ib_async event-driven programming, market data subscriptions, order execution with bracket orders, historical data with pacing rules, reconnection resilience, IBC automation, and Windows production deployment. TRIGGER WHEN: building, implementing, writing, coding, creating, optimizing, or debugging IB trading systems with Python. DO NOT TRIGGER WHEN: the task is outside the specific scope of this component.

Interactive Brokers Algotrading

Knowledge base for building production-grade algorithmic trading systems with Interactive Brokers TWS API and ib_async in Python.

When to Use

• Connecting to TWS or IB Gateway with ib_async
• Implementing real-time market data subscriptions
• Designing order execution logic (bracket orders, order lifecycle)
• Handling pacing violations for historical data requests
• Building reconnection resilience for 24/7 production bots
• Deploying IB trading systems on Windows with IBC + Task Scheduler
• Debugging TWS API error codes (110, 135, 162, 200, 201, 354, 366, 1100-1102)
• Translating FX/metals symbols into the right contract type (spot vs CFD) for the account's entity
• Sizing orders from prices without the NaN/0.0/unit/tick traps that silently mis-size or reject
• Diagnosing a bot that looks connected but is dead (zombie isConnected(), reconnect supervisor silently stopped)
• Event handlers that "never fire" (eventkit swallows listener exceptions; wrong handler arity)
• Bracket TIF design: GTC children, positions left naked overnight, orphaned SL/TP on flat contracts
• Historical FX bars off the bar grid (session-reopen stub bars) corrupting replays/bootstraps
• Daily market open/close events swallowed by a dedup layer (TTL resonance)

Quick Start

For 80% of use cases, start with:

1. Connection: IB Gateway (headless, lower resources) + IBC (automated lifecycle)
2. Library: pip install ib_async (asyncio-native successor to ib_insync)
3. Data: reqRealTimeBars for live 5-sec bars with local aggregation
4. Orders: Bracket orders with transmit=False pattern
5. Resilience: disconnectedEvent + exponential backoff reconnection
6. Deployment: IBC + Windows Task Scheduler for auto-restart

Then harden incrementally:

• Missing fills on reconnect -- add reqExecutions() reconciliation
• Pacing violations -- add asyncio.Semaphore throttled request queue
• Overnight crashes -- add IBC auto-restart + heartbeat monitoring
• State drift -- add periodic position/order reconciliation via reqPositions()
• Orders FIRED but never live, mis-sized, or rejected with no error -- read venue-boundary-failure-modes.md (async rejection ingress, tick conformance, FX-as-CFD routing, NaN-safe sizing)
• Connected-looking but dead for hours -- replace isConnected() gates with active probes (reqCurrentTimeAsync + timeout), defensive disconnect() after failed attempts, escalation on supervisor silence (reconnection-resilience.md)
• Handlers that never fire / invisible listener crashes -- contract-test handler signatures, route eventkit/ib_async std loggers to your sink (event-driven-data.md)
• Positions naked overnight or orphaned SL/TP -- GTC bracket children + residual-child reaper on position-closed (order-execution.md)
• Replay/bootstrap corrupted by bars live never saw -- drop off-grid session stub bars, chronology-guard state transitions (event-driven-data.md)

Reference Materials

• tws-api-architecture.md -- TWS API 10.45, Gateway vs TWS, Client Portal, ib_async setup, clientId strategy, official docs
• event-driven-data.md -- reqMktData, reqRealTimeBars, reqTickByTickData, keepUpToDate, OHLCV construction, pacing violations, historical data, session-reopen stub bars (off-grid drop + top-up, replay chronology guards), eventkit listener contracts (swallowed exceptions, handler arity, std-logger routing), dedup TTL resonance with daily market events
• order-execution.md -- order types, bracket orders (GTC children + residual-child reaper, phantom transient Cancelled), lifecycle states, execDetails monitoring, race conditions, error codes, compliance-201 non-overridability, whatIf probing, per-clientId order visibility
• reconnection-resilience.md -- daily reset, IBC automation, reconnect patterns, the isConnected() zombie blind spot (active liveness probes, defensive disconnect, decorrelated recovery layers), Gateway-log forensics, multi-client snapshot hygiene, the silent-failure review signature, heartbeat, Windows deployment, community resources
• venue-boundary-failure-modes.md -- the silent-failure layer: async rejection ingress (errorEvent -> lifecycle), price/tick conformance (110/135, minTick from ContractDetails), FX-as-CFD routing for retail EU entities (201/200/2127/366), CFD venue params vs spot (minTick, minSize-is-precision), data-contract vs order-contract split, and NaN-safe sizing with canonical lot units and conversion-rate fallback

Key Decision Points

Decision	Default	Upgrade When
Connection target	IB Gateway	Need visual debugging -- TWS
Python library	ib_async	Need same-day new features -- ibapi
Live data	reqRealTimeBars (5s bars)	Need tick precision -- reqTickByTickData (max 3)
Chart data	keepUpToDate	Network-sensitive env -- reqRealTimeBars + aggregation
Historical data	reqHistoricalData + throttle	Bulk backfill -- chunked requests with Semaphore
Order type	Bracket (parent+TP+SL)	Need trailing: TRAIL. Need algo: Adaptive
Lifecycle mgmt	IBC + Task Scheduler	Docker available -- gnzsnz/ib-gateway-docker
whatToShow	TRADES	Forex: MIDPOINT. Backtesting: ADJUSTED_LAST