This paper introduces a novel approach to studying market microstructure by analyzing linear trends in tick-level bid-ask data. The author derives a trend arbitrage inequality that relates the bid-ask spread to the absolute trend slope, market reaction time, stock-specific latency, and transaction costs. Using tick data from 33 stocks across five European exchanges, the paper identifies trends via a linear fitting algorithm with a chi-squared trigger. The key empirical finding is that the distribution of the response time (τ_resp = spread/|trend| - average update time) is log-normal, with a sharp front edge that defines the market reaction time τ_M. This τ_M is found to be exchange-specific, ranging from about 1.2 seconds in Frankfurt to 5.5 seconds in Madrid, and is interpreted as the latency of the exchange system. The analysis also shows that the product of trend and update time (μ·τ_u) is normally distributed with a standard deviation such that three sigma lies within the tick size, indicating that price updates occur frequently enough to prevent trend arbitrage. When tick size is subtracted from the spread, the measured τ_M values across exchanges converge to a narrow range of 0.6-1.2 seconds, suggesting that tick size absorbs much of the stock-specific latency. The paper further identifies three investor groups (fast, medium, slow) with different fee and latency characteristics. A simulation using a random walk model reproduces the normal distribution of μ·τ_u, suggesting a duality between trend-based and return-based descriptions of price dynamics. Overall, the study provides a microstructure framework linking trend, spread, latency, and tick size, with implications for understanding market efficiency and arbitrage opportunities.