Guides / Cookbooks

Cookbooks · Screening and digests · Beginner · about 10 min

Rank a watchlist by probability

Ask the same three questions about every symbol on a list, send one request per symbol in parallel, and sort the list by the answers.

1. When to use it

   You keep a list of 5 to 100 names and want a ranking after the close or before the open: which ones lean up for the week, and which ones have the widest monthly range. Each symbol is independent, so the requests can run side by side. For the same thing on a schedule with no code, see Send a pre-open digest.

2. The request

   One request per symbol. Only state.symbol changes; the server fills in that symbol's latest completed session.

   {
     "model": "ora-fin1",
     "state": {
       "symbol": "AAPL"
     },
     "questions": {
       "rally5": {
         "type": "yes_no",
         "instructions": "Will it rally at least 5% at some point in the next month?"
       },
       "drop5": {
         "type": "yes_no",
         "instructions": "Will it drop at least 5% at some point in the next month?"
       },
       "up_week": {
         "type": "yes_no",
         "instructions": "Will the stock be higher in a week?"
       }
     }
   }

   Open this in the Playground

3. The real response

   For AAPL:

   {
     "model": "ora-fin1",
     "answers": {
       "rally5": {
         "type": "yes_no",
         "yes_no": 0.446
       },
       "drop5": {
         "type": "yes_no",
         "yes_no": 0.365
       },
       "up_week": {
         "type": "yes_no",
         "yes_no": 0.59
       }
     },
     "usage": {
       "input_tokens": 362,
       "output_tokens": 12
     },
     "context": {
       "source": "server",
       "symbol": "AAPL",
       "as_of": "2026-10-30"
     },
     "version": "ora-fin-0.1.1"
   }

   All eight symbols, ranked by up_week (session of 2026-10-30):

   Symbol	up_week	rally5	drop5	rally5 − drop5
   XLE	0.592	0.491	0.491	+0.000
   AAPL	0.59	0.446	0.365	+0.081
   AMD	0.577	0.836	0.763	+0.073
   MSFT	0.574	0.495	0.422	+0.073
   QQQ	0.561	0.501	0.443	+0.058
   NVDA	0.527	0.616	0.523	+0.093
   AMZN	0.493	0.536	0.462	+0.074
   TSLA	0.487	0.687	0.638	+0.049

4. The code

   The single request in your language:

   cURLPythonJavaScriptPHPGoJavaC#RubyCopy

   Requires curl 7.76+. Save as run.sh, then: bash run.sh

   curl -sS https://oramodel.com/v1/systemone \
     -H "Authorization: Bearer $JEV_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
     "model": "ora-fin1",
     "state": {
       "symbol": "AAPL"
     },
     "questions": {
       "rally5": {
         "type": "yes_no",
         "instructions": "Will it rally at least 5% at some point in the next month?"
       },
       "drop5": {
         "type": "yes_no",
         "instructions": "Will it drop at least 5% at some point in the next month?"
       },
       "up_week": {
         "type": "yes_no",
         "instructions": "Will the stock be higher in a week?"
       }
     }
   }'

   The whole list, four requests at a time, ranked:

   import os
   import requests
   
   URL = "https://oramodel.com/v1/systemone"
   HEADERS = {"Authorization": f"Bearer {os.environ['JEV_API_KEY']}"}
   from concurrent.futures import ThreadPoolExecutor
   
   SYMBOLS = ["AAPL","MSFT","NVDA","TSLA","AMZN","AMD","QQQ","XLE"]
   QUESTIONS = {
       "rally5": {
           "type": "yes_no",
           "instructions": "Will it rally at least 5% at some point in the next month?"
       },
       "drop5": {
           "type": "yes_no",
           "instructions": "Will it drop at least 5% at some point in the next month?"
       },
       "up_week": {
           "type": "yes_no",
           "instructions": "Will the stock be higher in a week?"
       }
   }
   
   def ask(symbol):
       r = requests.post(URL, headers=HEADERS, timeout=30,
                         json={"model": "ora-fin1", "state": {"symbol": symbol}, "questions": QUESTIONS})
       if r.status_code == 422:          # symbol not covered: skip it
           return symbol, None
       r.raise_for_status()
       return symbol, {k: v["yes_no"] for k, v in r.json()["answers"].items()}
   
   with ThreadPoolExecutor(max_workers=4) as pool:
       rows = [(s, a) for s, a in pool.map(ask, SYMBOLS) if a]
   
   for s, a in sorted(rows, key=lambda r: r[1]["up_week"], reverse=True):
       print(f"{s:5} up_week {a['up_week']:.3f}  rally5 {a['rally5']:.3f}  drop5 {a['drop5']:.3f}  skew {a['rally5'] - a['drop5']:+.3f}")

   In JavaScript (Node 18+):

   const SYMBOLS = ["AAPL","MSFT","NVDA","TSLA","AMZN","AMD","QQQ","XLE"];
   const QUESTIONS = {
     "rally5": {
       "type": "yes_no",
       "instructions": "Will it rally at least 5% at some point in the next month?"
     },
     "drop5": {
       "type": "yes_no",
       "instructions": "Will it drop at least 5% at some point in the next month?"
     },
     "up_week": {
       "type": "yes_no",
       "instructions": "Will the stock be higher in a week?"
     }
   };
   
   async function ask(symbol) {
     const res = await fetch("https://oramodel.com/v1/systemone", {
       method: "POST",
       headers: { Authorization: `Bearer ${process.env.JEV_API_KEY}`, "Content-Type": "application/json" },
       body: JSON.stringify({ model: "ora-fin1", state: { symbol }, questions: QUESTIONS }),
     });
     if (res.status === 422) return null;               // not covered
     if (!res.ok) throw new Error(`${symbol}: ${res.status}`);
     const { answers } = await res.json();
     return { symbol, ...Object.fromEntries(Object.entries(answers).map(([k, v]) => [k, v.yes_no])) };
   }
   
   const rows = (await Promise.all(SYMBOLS.map(ask))).filter(Boolean);
   rows.sort((a, b) => b.up_week - a.up_week);
   console.table(rows);

   Keep the pool small: limits are per organization and shared with the console. On a 429, wait Retry-After seconds (Handle errors and retries).

5. Act on it

   A ranking is relative. On this session the top up_week is XLE at 0.592 and the bottom is TSLA at 0.487: nothing is far from a coin flip on the week, so a sensible screen acts on none of them and just watches the top of the list. Decide the line first (for example up_week >= 0.62), then rank what clears it. Read probabilities and choose thresholds shows how.

   The other columns tell a different story. AMD has the highest odds of both a 5% rally and a 5% drop (84% and 76%): a wide range rather than a direction. That case has its own recipe, Detect two-way risk.

6. Variations

   • Rank by rally5 − drop5 to see which way the month's range leans.
   • Add a rule such as {"type": "yes_no", "ask": "C > sma(C,200)"} and keep only names in an uptrend (Trend filter plus entry trigger).
   • Put all the questions you need into the one request per symbol, not one request per question (Ask many questions in one request).
   • Run it every day without code as a workflow with a digest (Send a pre-open digest).

Information, not investment advice. Every number on this page is a real response for the session shown; answers change with each session.

What's next

• Detect two-way risk
• Send a pre-open digest