Pinnacle Data Corp CLC Database

The Pinnacle Data Corp CLC (Continuous Linked Contracts) Database is a commercial dataset of ratio-adjusted (back-adjusted) continuous futures contracts, splicing successive delivery months into a single continuous price series with roll discontinuities removed. It appears in this vault as the data source for Zhang Zohren Roberts 2019, which used the 50 most liquid contracts — 25 commodity, 11 equity-index, 5 fixed-income and 9 FX — over 2005-2019 to train the DQN/PG/A2C reinforcement-learning agents that produced the strongest positive, post-cost, out-of-sample RL-trading backtest tracked here. Continuous-contract construction is convenient for research but can introduce subtle look-ahead artefacts if the back-adjustment is mishandled, a caveat relevant to grading any backtest built on it.

Connections

Sources