Now from this list narrow it to two stocks

Important Note on the Request

It isn’t possible to guarantee that any stock will gain an additional 5% in a single day. What we can do is use historical patterns and probabilistic models to tilt the odds toward stocks that statistically have a higher chance of rising and whose expected one‑day return is in the desired range. The filters below are designed exactly for that.

Screening Filters

1. price: {'min': '3', 'max': '200'}`

   • Purpose: Focus on reasonably priced, liquid stocks and avoid ultra‑low‑priced penny stocks or very high‑priced outliers.
   • Rationale:
     • Stocks under $3 are often illiquid, highly speculative, and prone to extreme, random moves that can distort any short‑term prediction model.
     • Very high‑priced stocks (above $200) often move in dollar terms more than in percentage terms; a 5% intraday move on a $500 stock is large and rarer.
     • Keeping the range between $3 and $200 helps target stocks where a 4–5% intraday move is more common and where prediction models tend to be more stable.

2. price_change_pct: {'min': '-1', 'max': '6'}`

   • Purpose: Select stocks that have moved modestly today, but haven’t already “spent” most of their potential upside.
   • Rationale:
     • Lower bound -1%: Excludes stocks that are plunging hard; a deep intraday loss can signal bad news or abnormal risk, lowering the practical odds of a clean 5% upside from here.
     • Upper bound 6%: Excludes stocks that are already up sharply. If a stock is already up, say, 10% today, the probability of gaining an additional 5% is typically lower because some of the move may have already occurred.
     • Keeping moves between -1% and +6% balances: the stock is active, but not exhausted.

3. region: ['United States']

   • Purpose: Limit the universe to U.S.-listed companies.
   • Rationale:
     • U.S. markets (NYSE, NASDAQ, AMEX) are highly liquid and well‑covered, which improves data quality and model reliability for short‑term probability and return estimates.
     • It also aligns with many users’ preference for regulatory transparency and trading convenience.

4. list_exchange: ['XNYS', 'XNAS', 'XASE']

   • Purpose: Restrict results to major U.S. exchanges: NYSE, NASDAQ, and NYSE American.
   • Rationale:
     • These exchanges enforce listing standards (financial, reporting, liquidity) that help filter out the riskiest and least transparent names.
     • Short‑term prediction models usually perform better on actively traded, well‑listed stocks rather than on obscure OTC or pink‑sheet securities.

5. one_day_rise_prob: {'min': '55'}`

   • Purpose: Require a probability > 55% that the stock closes higher than its current price today, based on a quantitative model.
   • Rationale:
     • A 50% rise probability is essentially a coin flip. Setting the bar at 55% means we only keep stocks where the model sees a meaningfully above‑average chance of going up today.
     • This directly reflects your request for “potential of gaining” today, by focusing on names where the short‑term odds are tilted towards a positive move.

6. one_day_predict_return: {'min': '4', 'max': '20'}`

   • Purpose: Target stocks whose expected one‑day return is between +4% and +20%.
   • Rationale:
     • Lower bound 4%: Aligns with your “additional 5%” goal by ensuring we are looking at stocks where the model expects a sizable positive move close to that range (4–5%+).
     • Upper bound 20%: Avoids extreme outliers where the model predicts very large returns that often come with extreme risk or low reliability.
     • This range focuses on stocks whose forecasted move is substantial enough to matter, but not so extreme as to be purely speculative.

Why Results Match Your Intent

• The rise probability filter (≥55%) ensures we’re not just picking random volatile stocks, but ones with a statistically higher chance of going up today.
• The predicted return filter (4–20%) is directly tied to your “additional 5 percent” objective, ensuring the expected move is in that neighborhood or higher.
• The price and price_change_pct filters help avoid both penny‑stock noise and already‑overextended names, improving the realism of a further 4–5% intraday gain.
• Limiting to major U.S. exchanges enhances liquidity, data quality, and the reliability of the short‑term models used to estimate these probabilities and returns.

Taken together, these filters don’t guarantee a 5% gain, but they systematically select stocks where a meaningful positive move today is statistically more plausible, from a more stable and liquid subset of the market.