How momentum strategies pick stocks comes down to one measurable observation: stocks that have outperformed recently tend to keep outperforming in the near term, and stocks that have underperformed tend to keep underperforming. That empirical pattern — called price momentum — is one of the most researched factors in equity finance, and it forms the backbone of several of VelaDeck's AI-curated monthly strategies.
How Momentum Strategies Score and Rank Stocks
A momentum model starts with a universe — all US equities above a certain liquidity threshold, for example — and gives each stock a score. The core of that score is the stock's trailing return over a defined window, measured relative to its peers in the same universe.
A stock that has returned substantially more than the median name over the past several months carries a strong positive score. One that lagged significantly carries a weak score. The model sorts all candidates and selects the top N — twenty names for a twenty-stock strategy, fifteen for a fifteen-stock one.
AI models layer additional signals on top of raw price momentum: earnings estimate revisions, quality of earnings, valuation relative to history. Price momentum tells the model what the market has already priced in about improving or deteriorating fundamentals; the other factors help filter out names where the momentum is on borrowed time.
From Score to Portfolio Composition
Once the model has its ranked list, it publishes a composition: these tickers, in these approximate weights. That composition is the strategy's current state.
The next month, the model re-runs the entire scoring pass. Some names score higher and enter the portfolio for the first time — a BUY signal. Others drop below the cut and exit — a SELL signal. A few stay in but shift weight — a REWEIGHT signal. Some months nothing changes. Others, after an earnings season or a macro event, turnover spikes and a meaningful portion of the portfolio rotates.
What the model does not do: it does not time individual trades within the month, react to intraday moves, or predict precise price targets. The composition is a monthly snapshot, not a continuous trading signal.
The Diff: Where BUY, SELL, and REWEIGHT Signals Come From
VelaDeck's ingester pulls each strategy's composition daily and diffs it against yesterday's snapshot. Every change in that diff becomes a signal.
A ticker that appeared in today's composition but not yesterday's becomes a BUY. A ticker present yesterday but absent today becomes a SELL. A ticker in both but with a materially different weight becomes a REWEIGHT. Those signals feed a planned-order engine. For each user that has assigned USD to a strategy, VelaDeck generates one planned order per signal, sized by the user's notional. SELL orders use your current position quantity so VelaDeck never over-sells. BUY orders use notional (fractional shares are supported).
Every order gets a deterministic `client_order_id` derived from the signal. If the executor fires twice on the same signal, Alpaca deduplicates it and VelaDeck marks the row as `sent_dedupe` rather than double-buying.
See [how VelaDeck automates monthly rebalancing on Alpaca](/blog/how-veladeck-automates-alpaca-rebalancing) for the full pipeline from composition to filled order.
Monthly Cadence Concentrates the Trading Activity
Because the composition update happens monthly, the bulk of trading activity is front-loaded to one day per month. Between rebalances, a momentum portfolio is usually quiet — no signals to execute, no decisions to make.
This cadence matters for transaction cost. Commission-free brokers like Alpaca have eliminated per-trade fees, but market impact still applies for large notional amounts. Monthly rebalancing keeps the number of round-trips low relative to a strategy that repositioned weekly.
VelaDeck's [Mid-Cap Movers and Beat the S&P 500 strategies](/strategies) are both structured around this monthly update rhythm. Historical performance for each is available on the strategies page — but past results are historical and simulated, not a prediction of future returns.
Execution Risk Is Real and Containable
The gap between a strategy's modeled composition and your actual holdings is where execution risk lives. A BUY order sent at 8:15 ET fills at market open — not at the previous close. If the stock gaps up overnight, your cost basis is higher than the model's assumed entry.
For portfolios that rebalance monthly, that one-day fill price difference is usually small relative to the return generated over months. But it compounds if you consistently miss rebalances or execute late. VelaDeck automates the submission to minimize the gap.
Live trading is off by default — everything runs against a paper Alpaca account until you explicitly opt in. Before going live, read [from paper to live trading, safely](/blog/paper-to-live-trading-safely) for the recommended sequence. VelaDeck is not an investment advisor or broker-dealer, and nothing here is a recommendation to buy or sell any security. Your money stays in your own Alpaca account — VelaDeck does not custody funds. Historical and simulated performance does not predict future results.
What exactly is price momentum in equity selection?
Price momentum is the trailing return of a stock relative to its peers over a defined window — commonly several months. A stock with strong positive momentum has outperformed its universe recently. AI models use it as one signal alongside earnings, quality, and valuation factors, not in isolation.
How often does a momentum strategy replace its holdings?
Monthly for most VelaDeck strategies. Some months the composition barely changes; others, following an earnings season or macro event, several positions rotate out. The exact turnover depends on the strategy and the market environment at the time of the scoring pass.
Does a momentum signal mean I need to place orders manually?
No. VelaDeck submits market orders before market open on the day a signal is confirmed. You do not manage the timing yourself — the system handles submission, and the fill is recorded with the actual average price Alpaca returned.
Do I need to pick the stocks myself?
No. The strategy selects and weights the stocks. What you control: which strategy or combination of strategies you allocate to, how much USD you assign to each, and whether to set a monthly cap. You can model different allocations in paper mode on the [strategies page](/strategies) before committing real capital.
Can momentum strategies underperform for extended periods?
Yes. Momentum as a factor tends to underperform during sharp market reversals and early recovery phases. Allocating across strategies with different underlying signals — value-oriented, tech-tilted, mid-cap growth — reduces single-factor exposure. [Sign up](/signup) to start with a paper Alpaca account and explore blended allocations at no cost.