Quantifying Slippage in Vectorbt: Tick Data
What if your backtest results are lying to you because you cannot see the slippage hidden inside the bars? Accurate Vectorbt slippage modeling requires ...
What if your backtest results are lying to you because you cannot see the slippage hidden inside the bars? Accurate Vectorbt slippage modeling requires more than just a flat fee; it demands a granular view of market microstructure that standard OHLCV data simply cannot provide. The core difference lies in information density. Tick data is the raw, unfiltered stream of every trade execution and quote change, preserving the exact sequence of events. In contrast, OHLC data compresses all activity within a fixed time interval into four values: Open, High, Low, and Close. When you aggregate thousands of individual events into a single bar, you lose the sequence of trades, the distribution of volume across price levels, and the precise timing of price movements. Key fact: A single day of tick data for one liquid instrument can easily run into hundreds of megabytes, whereas the same period compressed into one-minute OHLC bars might only require a few kilobytes. This compression is lossy. Once you collapse thousands of events into four price points, you cannot reconstruct the path the price took to get there.