You guys focus on traffic, plagiarism, can't you have some bottom line 😅

This Kronos with a 93% win rate, don't you verify it when plagiarizing?
First, the "93% win rate" is completely false advertising. The paper clearly states: the RankIC indicator for price sequence prediction improves by 93% over leading time series baseline models, and by 87% over non-pretrained baselines.
RankIC is an academic ranking correlation metric, not an actual trading win rate or prediction accuracy.
Moreover, this is just a research-based foundational model, not an out-of-the-box trading magic tool. The repository README and the paper both explicitly disclaim: the model output is not "pure alpha," and the demonstration strategy is for reference only. Real trading must consider slippage, transaction costs, risk management, position control, etc.
Not to mention whether there is a stable 93% win rate model—if there were, why would the author open source it for you to get rich? Unbelievable.