Model wrong
pavybez opened this issue · 11 comments
Nice implementation but AS model is wrong. You need to multiply the vol% in config with current price to get sigma.
Thank you for pointing out, could you give a more detailed explaination?
Lets say you configure sigma as 0.03 that means that volatility is 3% of current spot price. Use that value to calculate spreads.
Ah, I see. I'll look into it. Will update here when it is fixed
In the meantime, feel free to send a pull request if there is any fixes or improvements you'd like to add.
A couple more suggestions please:
- Get the mark price ".BXBT" as well cause sometimes on Bitmex there is a huge gap between mid price and mark price and incorporate it into the model.
- Compute volatility online as mark price changes. See here: https://core.ac.uk/download/pdf/52391988.pdf and use the fact that variance is proportional to time.
- Compute OBV signal cause on Bitmex, you will notice that when bid or ask size grows out of proportion it moves in the other direction significantly.
- Incorporate max inventory position and quote only on one side if current inventory (long or short) exceeds max inventory.
- Have a mechanism to close positions if they hit a stop loss criteria.
If you do these you strategy will become better. Good luck.
Thank you, these are very valuable inputs. I'll add them into the product plan. Looking forward to your suggestions & critics when these are done
Noted, will look into it
sorry to bother you.
I don't understand this issue.
you mean to calculate sigma with leverage? @pavybez @hello2all
thank you~
So sorry to bother you again. I am not major in finance, the knowledge base is a little bit poor.
Now I know why I should convert sigma in % to usdt, but I dont know how to do it? There are some method to calculate sigma in %, but not mention how to convert it, mabe mutiply by mean value or orther else? If any paper mention it or any methods, please tell me.
thanks!
@pavybez