Markov Method 2.0 without Fable 5?

21d • Ask

I'm writing from the US where Fable 5 has been "banned" by the U.S. government! Therefore, the latest model I have access to is Opus 4.8. I had previously run the Markov 1.0 code and was able to successfully evaluate tickers within Claude code. I was also able to copy and paste the Pine script and properly create the indicator in TradingView. (AWESOME!!!! THANK YOU SO MUCH!!!)

My questions now:

1) It seems that re-running the script in Opus 4.8 (instead of Fable 5) has delivered some questionable results. eg: See screenshot for the 'FIX 1- Stride sampling results'. According to this table, the 'Stride-sampled (true) 'BULL->BULL' is only 0.067 when evaluating SPY... That seems remarkably low. Are the values in this table (screenshot) the same values that are created when this script is run with Fable 5?

2) The Pinescript did not automatically get generated to apply the Markov 2.0 indicator to TradingView. I have the latest version of the TradingView desktop app installed on my Mac. Is this limitation driven by the fact that I am using Opus 4.8 instead of Fable 5? And if so, would it be possible for you to copy the Pinescript for the Markov 2.0 indicator and add it to the YouTube prompts folder?

Thank you again so much for all this amazing work for free! I am getting started way too late with Claude, but you are really helping me make up for lost time! Can't thank you enough!

8 comments

Markov Method 2.0 without Fable 5?