The strategy seeks to harvest short term regime persistence in large cap U.S. equities by classifying recent price momentum into latent states using a 3 state Gaussian Hidden Markov Model. The trading universe is the 10 highest market capitalization stocks available from fundamentals, refreshed as constituents change. For each security, the signal input is a rolling window of 1 bar rate of change computed on intraday consolidated bars (default 5 minute bars) over roughly 150 observations. The model is refit monthly at the start of each month using the latest window to adapt to evolving return distributions. Once the model is trained, every new consolidated bar generates a directional forecast by comparing the posterior probability of an up regime versus a down regime; the strategy goes long when the highest probability state corresponds to the most bullish regime and goes short otherwise. Positions are sized equally across active names via an equal weight portfolio, with portfolio checks on a 5 minute cadence and no forced liquidation or rebalance on each insight change, implying positions can persist until new signals gradually shift the target weights. The edge is statistical regime detection on high frequency momentum, aiming to exploit short horizon trend continuation versus reversal between inferred states.