Maximization with a user-defined function

[sabrivaran]

Hello, I am working on a simple ema crossover strategy in a currency USDTRY and I come up with the "problem" that:
When optimizing my strategy with maximize='Equity Final [$]' ; equity is calculated in TRY not in USD.
Hence I need your help for defining an optimization function which will be based on the value Equity/ data.Close, so that I think it will be in USD. Defining such a function is easy probably but I could not figure it out how it will work with the strategy.
thanks from now.
I am definitely new to Python and Backtesting so any idea, help or recommendation is really valuable for me
Here is my code:

from backtesting import Backtest, Strategy
from backtesting.lib import crossover, barssince
import pandas as pd
import tulipy as tu 

datafile = """....\\USD_TRYinvesting 01.11.2019-15.04.2021.csv"""
data=pd.read_csv(datafile, index_col = 'Tarih',decimal=',')

data.index = pd.to_datetime(data.index, dayfirst=True)
data.columns = ["Close","Open","High","Low","Volume"]             
data=data.sort_index()
class SmaCross(Strategy):

    ema_no1=9
    ema_no2=20
    
    def init(self):        
        price = self.data.Close
        self.ema1 = self.I(tu.ema, price, self.ema_no1)
        self.ema2 = self.I(tu.ema, price, self.ema_no2)
        
    def next(self):
        if barssince(self.data.High ==5.7525 and self.data.Close==5.7124) > self.ema_no2:       
            if crossover(self.ema1, self.ema2):
               # self.position.close()
                self.buy()
            elif crossover(self.ema2, self.ema1):
               # self.position.close()
                self.sell()


bt = Backtest(data, SmaCross, commission=0.0004,cash=1000,
                exclusive_orders=True)

stats, heatmap = bt.optimize(ema_no1=range(5,15, 1),
                    ema_no2=range(10,50,1),
                    maximize='Equity Final [$]',
                    constraint=lambda param: param.ema_no1 < param.ema_no2,
                    return_heatmap=True)

bt.plot()
mytrades=stats._trades
mytrades.to_csv("trades3.csv")

hm = heatmap.groupby(['ema_no1', 'ema_no2']).mean().unstack()
from backtesting.lib import plot_heatmaps
plot_heatmaps(heatmap, agg='mean')

[kernc]

maximize= takes a function. Try:

final_price_in_try = data.Close.iloc[-1]

stats, heatmap = bt.optimize(..., maximize=lambda stats: stats['Equity Final [$]'] / final_price_in_try)

[sabrivaran]

Thank you for your quick reply but unfortunately that is not giving the result I am looking for. It seems to me that with your maximization function Equity variable is just divided by the dataset's last closing price.

What I need is; dividing Equity during the strategy by the every bar's close price

[kernc]

my strategy with maximize='Equity Final [$]' ; equity is calculated in TRY not in USD

That's what USDTRY currency is. If you want to trade the inverse (and have stats based in USD), just feed the strategy with:

data[['Open', 'High', 'Low', 'Close']] = 1 / data[['Open', 'High', 'Low', 'Close']]