IEEE BIG DATA CUP 2026 / TASK 2
Market-neutral decisions.
Measured round by round.
Select two stocks, choose LS, SL, or F, and compare cumulative returns with the Codex baseline. All schedule times are in New York.
10 stocks12 data releases14–15 October 2026
Connecting to release status…
01 / POINT-IN-TIME INPUTS
Data releases ↗
One price row and one English news summary per ticker. Summaries contain 30–50 tokens.
| Round | New York release | News window | Action due | Files |
|---|
News windows include the start and exclude the end. Missing data is flagged; incomplete coverage is not treated as no news.
02 / YOUR NEXT DECISION
Submit a pair action ↗
Continue on Hugging Face. Stock selections and actions remain administrator-private.
The static Space currently shows submission records only. The private submission service is not yet enabled.
03 / SHARED PERFORMANCE
Cumulative returns
Live results disclose names and cumulative returns only. Final ranks appear after 16 October, 10:30 New York time.
Open the main website CR ranking ↗
| Final rank | Team / reference | Cumulative return |
|---|---|---|
| No verified cumulative returns have been published. | ||
Return paths
The final chart will appear after verified terminal settlement. No illustrative returns are presented as live results.
04 / TEST PROTOCOL
How the rounds work
1. Release and decide
Round 1 at 14 October 09:30 publishes data only. Releases 2–12 on 14–15 October launch both data collection/publication and settlement of the preceding round.
Releases occur at 09:30, 10:30, 11:30, 12:30, 14:30, and 15:30 each day. Intraday action deadlines are five minutes before the next release.
2. Settle consistently
At each subsequent event, evaluate the previous accepted pair action using the fixed valuation and scoring rules. Missing valid actions are scored as F. Compound completed-round returns.
The 14 October 15:30 action is due on 15 October at 09:25. The final action is due on 16 October at 08:00.
3. Seal the baseline
Codex chooses two stocks and LS, SL, or F from each round’s available data. Its decision is recorded privately before the following release, then scored with the same rules as every team.
Neither team actions nor baseline decisions are disclosed on the live board.
4. Finish the test
16 October 09:30 runs settlement only, using internal terminal prices. There is no thirteenth participant data release.
At 10:30, publish final rankings and cumulative-return curves after terminal settlement completes. The Codex baseline is a reference, not a competing team.
This is a competition simulation, not a brokerage execution service. Position sizing, valuation, execution timing, and transaction costs must be fixed in the verified scorer before trading opens.