data(leaderboard): publish the 4 remaining LLM entries (GPT-5.5, Gemini, Qwen, DeepSeek)#75
Open
FlyM1ss wants to merge 2 commits into
Open
data(leaderboard): publish the 4 remaining LLM entries (GPT-5.5, Gemini, Qwen, DeepSeek)#75FlyM1ss wants to merge 2 commits into
FlyM1ss wants to merge 2 commits into
Conversation
Adds full 161-step contest-window runs (2026-04-15→05-15) for three CommonStack-gateway LLM entries to the seed DB, each 100% LLM-decided (161/161 calls) — clearing the H6 95%-coverage guard (PR #74): GPT-5.5 +1.45% Sharpe 1.56 239 trades ~$13.89 Gemini 3.1 Pro Prev +2.32% Sharpe 2.03 267 trades ~$11.26 Qwen3.7 Plus +2.49% Sharpe 2.19 205 trades ~$1.59 Brings the board to 5 of 6 LLM models (Claude Haiku 4.5 + Sonnet 4.6 already live via #72/#73). DeepSeek V4 Pro to follow in a separate PR (slow reasoning model, still running). Snapshot taken via VACUUM INTO so all WAL-committed runs are captured in the committed file. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014t1xHNTNh5mybdjMQAvQcG
DeepSeek V4 Pro finished its full 161-step contest-window run, 100% LLM-decided (161/161 calls), clearing the H6 95%-coverage guard: DeepSeek V4 Pro +7.49% Sharpe 5.01 277 trades ~$0.76 It tops the LLM field and is the only agent to beat the passive baselines. Completes the CommonStack LLM set alongside GPT-5.5, Gemini 3.1 Pro, and Qwen3.7 Plus in this PR; all six leaderboard LLM models now have runs. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014t1xHNTNh5mybdjMQAvQcG
FlyM1ss
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds full 161-step contest-window runs (2026-04-15 → 2026-05-15) for three CommonStack-gateway LLM entries to the committed seed DB. Each is 100% LLM-decided (161/161 calls) and clears the H6 95%-coverage guard from #74 — no rule-based fallback masquerading as a model.
This brings the leaderboard to 5 of 6 LLM models (Claude Haiku 4.5 + Sonnet 4.6 already live via #72/#73).
Notes
dashboard/storage/data/backtest.dbchanges (additive: 3 new runs + their equity curves), mirroring feat(leaderboard): publish Claude Haiku 4.5 + Sonnet 4.6 contest entries #72.VACUUM INTOso all WAL-committed runs are captured in the committed file.leaderboard.jsonalready lists all six LLM entries, so no config change is needed — these entries simply needed cached runs.🤖 Generated with Claude Code