Tick level market data is the most granular form of financial data available. It records every individual trade and quote update as a separate event, in the exact order it occurred, with timestamps precise to the millisecond or microsecond. Unlike bar or candle data that summarizes activity over fixed intervals, tick data preserves every market event without compression or aggregation. Algorithmic trading systems rely on tick data because the signals they act on — order flow shifts, spread changes, and liquidity fluctuations — exist only at this resolution.
A tick is a single market event. It can be a completed trade between a buyer and a seller, a change to the best available bid or ask price, or an update to the volume available at a given price level. Each tick is recorded with a precise timestamp, a price, and typically additional fields such as trade size, exchange venue, and the prevailing bid-ask spread at the moment of the event.
Ticks do not arrive at regular intervals. A heavily traded stock like Microsoft or Apple can generate over 90,000 ticks in a single trading day. During earnings announcements or macroeconomic releases, thousands of ticks may arrive within a single second. During quiet periods, several seconds may pass with no updates at all. This irregular spacing reflects actual market behavior, not a limitation of the data. The entire U.S. equity market generates billions of ticks per day across all listed instruments.
A typical tick-level record contains the following fields:
Timestamp — Time of the event, precise to millisecond or microsecond
Price — Execution price (for trades) or quoted price (for bid/ask updates)
Size — Number of shares or contracts in the trade or quote
Side — Whether the event is a trade, bid update, or ask update
Exchange — The venue where the event occurred (e.g., NYSE, Nasdaq, BATS)
Bid price — Best available buying price at the time of the event
Ask price — Best available selling price at the time of the event
Bid size — Volume available at the best bid
Ask size — Volume available at the best ask
Not every dataset includes all of these fields. The minimum is a timestamp and a price. Professional and institutional datasets typically include the full set, along with condition codes indicating special trade types such as odd lots, block trades, or late reports.
Bar data (also called candle data or OHLCV data) summarizes trading activity over a fixed time interval — typically one minute, five minutes, or one day. Each bar reports the opening price, highest price, lowest price, closing price, and total volume for that interval.
Tick data differs from bar data in three important ways:
Resolution — A one-minute bar compresses hundreds or thousands of individual events into four prices and a volume count. Tick data retains every event. Two bars with identical OHLCV values can represent completely different market conditions, one steady and calm, the other volatile and reversing. Only tick data reveals the difference.
Sequence — Bar data does not record the order in which events occurred within the interval. Tick data preserves the exact sequence. This matters for any analysis that depends on how a price moved from point A to point B, not just that it moved.
Timing — Bars arrive at fixed intervals regardless of market activity. Ticks arrive whenever the market changes, producing dense clusters during active periods and gaps during quiet ones. This irregular timing is itself a signal, which reflects where market participants are concentrating their attention.
Algorithmic trading systems operate on signals that exist at very short time scales. Three areas depend directly on tick-level resolution:
Execution analysis — When an algorithm places an order, the quality of the fill depends on what happened in the milliseconds around that order. Did the spread widen just before execution? Did available liquidity disappear between order submission and matching? Tick data makes these questions measurable. Without it, execution quality assessment relies on approximation.
Backtesting accuracy — A strategy backtested on one-minute bars assumes that the prices in those bars were continuously available, which is often false. The high of the bar may have lasted for a fraction of a second. The close may have occurred on a single small trade. Tick-level backtesting reveals whether the prices a strategy claims to have traded at were actually accessible given the real order flow at the time.
Market microstructure research — How prices form, how information enters the market, how different participant types interact, and how market makers manage inventory all unfold at the tick level. Trying to study these dynamics with aggregated data is like trying to understand a conversation by looking only at the word count.
Trade ticks and quote ticks record different types of events:
Trade ticks record completed transactions. Two parties agreed on a price and exchanged shares or contracts. A trade tick shows where actual business happened and at what size.
Quote ticks record changes to standing orders in the market whenever someone posts, adjusts, or withdraws a bid or an ask. Each quote update reflects the evolving balance of supply and demand around transactions and shows how the market is positioned before, during, and after each trade.
Datasets that combine both are called TAQ (trades and quotes) data. TAQ data provides the most complete picture because each trade can be evaluated in the context of the prevailing quotes at that moment, making it possible to determine whether a fill was favorable, neutral, or adverse relative to available liquidity.
Two main distribution methods exist for tick data:
Consolidated feeds aggregate trades and quotes from all exchanges into a single stream. In U.S. equities, the Securities Information Processor (SIP) performs this function. Consolidated feeds provide a complete view of national best bid and offer (NBBO) across all venues, but introduce a small amount of additional latency due to the aggregation step.
Direct exchange feeds deliver data from a single exchange with lower latency. They bypass the SIP and provide venue-specific information faster, but require separate connections to each exchange. High-frequency and latency-sensitive strategies typically use direct feeds. Research and post-trade analysis often rely on consolidated feeds because they capture the full market in one stream.
The choice between consolidated and direct feeds depends on the use case. Latency-sensitive strategies need direct feeds. Comprehensive market analysis and regulatory reporting need consolidated feeds.
Tick data introduces requirements that aggregated data does not:
Storage — A single actively traded stock can produce gigabytes of tick data per year. A full market across all listed instruments requires infrastructure built for terabyte-scale datasets.
Timestamp precision — Events must be recorded to the millisecond or microsecond to preserve correct sequencing. When two events arrive within the same millisecond, their order can still affect analysis, particularly for strategies sensitive to queue position or information arrival.
Data quality — A single erroneous price inside a one minute bar may have little practical impact. In a tick stream, however, that same error can trigger false signals in an algorithm, distort spread calculations, or contaminate an entire backtest. For that reason, filtering and validating tick data by removing duplicate records, correcting out of sequence timestamps, and flagging anomalous prices is an essential step in any production workflow.
Processing speed — Algorithms consuming live tick data must process each event and generate decisions within microseconds. This demands specialized hardware, optimized code paths, and often co-located infrastructure near exchange matching engines.
Quantitative hedge funds, proprietary trading firms, algorithmic execution desks, academic researchers studying market microstructure, and financial regulators performing surveillance and market reconstruction all use tick data. Retail traders occasionally access tick data for research, though the storage and processing requirements make it primarily an institutional resource.
Tick data pricing varies widely by provider, asset class, and historical depth. U.S. equity tick data from major providers typically costs thousands of dollars per year for historical archives. Real-time direct exchange feeds involve exchange licensing fees in addition to data vendor costs.
Some brokers provide limited real-time tick data to account holders. The SEC's MIDAS system publishes aggregated tick-level statistics. Academic researchers may access tick data through university subscriptions to databases like TAQ from NYSE or LOBSTER for limit order book data. Comprehensive historical tick data with full market coverage is generally not available for free.
Common formats include CSV for simple flat-file storage, binary formats like Parquet and HDF5 for efficient compression and fast query performance, and PCAP for raw network packet captures of exchange feeds. The choice of format typically reflects the trade-off between human readability and processing efficiency.
Major U.S. equity tick data providers offer archives going back to the early 2000s, with some datasets extending into the 1990s. Data quality and completeness generally improve in more recent years as exchange reporting standards have tightened and electronic trading has become dominant.