I am developing a bot for Isolated Margin where I automatically add a “Stop Loss” position for every position to avoid liquidation.
Calculating the correct liquidation price for Isolated Margin is somewhat complicated due to the tiered system:
Tiered Leverage Function on Isolated Margin | Binance Support
The Query Isolated Margin Account Info in the API returns “liquidatePrice”:
https://binance-docs.github.io/apidocs/spot/en/#query-isolated-margin-account-info-user_data
Can I trust the value returned here for my Stop Loss position? I have done some trials, and the value I get is 99.9% equal to my calculations in all cases.
My questions are:
-
Can I trust the “liquidatePrice” as a basis for calculating the execution price for my Stop Loss positions?
-
Will Binance compensate an automatic liquidation if my API log shows that the API gave a wrong value, and I acted in good faith? I.e. I placed a Stop Loss position that was too low/high, based on the value I got from the Binance API.
I am aware that I must calculate in the interest on the loans, which will be added to the loans, a tiny fraction every hour.
To calculate the liquidation price, I solved the Margin Level equation in Maple (lazy me) for BTC/USDT Isolated Margin:
I then implemented the solution as a function in Python, to get the liquidation price:
.
def calculate_liquidation_price(USDT_balance, USDT_loan, BTC_balance, BTC_loan, margin_level):
a = USDT_loan * margin_level - USDT_balance
b = BTC_balance - BTC_loan * margin_level
if b == 0:
return sys.float_info.max
else:
return a / b
.
I tested this with USDT 100 on Isolated Margin BTC/USDT throught the Binance API for different values of Long and Short positions with x4 leverage for both BTC and USDT. The resulting “liquidatePrice” values that I got back from Binance matched very well the results from the Python function with margin_level 1.05, as it should.
.
In the cases where the formula gave a negative price (which indicates that there is no need for a Stop Loss), Binance returned “liquidatePrice: 0". I find the choice of zero for negative values to be sound, as returning a negativ liquidation price would just create confusion.
.
Because the returned “liquidatePrice” value from Binance matches the formula so well, it is tempting to simply trust it instead of doing my own math every time I need to place a Stop Loss position.
I did some work in Maple to understand the liquidation formula for a situation where I have loans both for USDT and BTC in an Isolated Margin account. x4 leverage for both.
This gives an interesting result: For small Long positions, there is no need to place a Stop Loss position.
Bottom line is that the values that Binance returned when I tried this with real positions on my BTC/USDT Isolated Margin account matched the theoretical results 100%.
