backtest.py - neatgrid - Division by zero

[ilaykim]

**Really weird, only managed to catch this on SOLUSDT on Neat Grid.

tried refreshing tick data, didn't work.
OHCLV - same issue
different json - same issue**

different symbol - works
short disabled - works
long disabled - DOESN'T WORK

python3 backtest.py D:\bot\psv57\configs\sol_655days.json -s SOLUSDT -b D:\bot\psv57\configs\backtest\sol_2M.hjson

using numba
using numba
using numba
fetching market_specific_settings...
linear perpetual

exchange binance
spot False
symbol SOLUSDT
market_type futures
starting_balance 1000
start_date 2022-05-09
end_date 2022-07-09
latency_simulation_ms 1000
base_dir backtests

2022-07-11T07:08:31 Loading cached tick data from backtests\binance\SOLUSDT\caches\2022-05-09_2022-07-09_ticks_cache.npy
{'config_name': 'neat_grid_SOLUSDT_655days',
'logging_level': 0,
'long': {'auto_unstuck_ema_dist': 0,
'auto_unstuck_wallet_exposure_threshold': 0,
'backwards_tp': True,
'ema_span_0': 120.62199728529626,
'ema_span_1': 34.76100179121808,
'enabled': True,
'eprice_exp_base': 1.1618608307844618,
'eqty_exp_base': 2.7802305422341758,
'grid_span': 0.17566515453113268,
'initial_eprice_ema_dist': -0.09535081461351516,
'initial_qty_pct': 0.014571840796302686,
'markup_range': 0.009421624486160773,
'max_n_entry_orders': 8,
'min_markup': 0.002,
'n_close_orders': 9,
'wallet_exposure_limit': 0.8},
'short': {'auto_unstuck_ema_dist': 0,
'auto_unstuck_wallet_exposure_threshold': 0,
'backwards_tp': True,
'ema_span_0': 4237.55904931995,
'ema_span_1': 9054.894470988955,
'enabled': True,
'eprice_exp_base': 1.161340927050355,
'eqty_exp_base': 3,
'grid_span': 0.4209059738318962,
'initial_eprice_ema_dist': -0.03917532734094498,
'initial_qty_pct': 0.04879303398638066,
'markup_range': 0.03261166232435841,
'max_n_entry_orders': 13,
'min_markup': 0.012619198687759918,
'n_close_orders': 17,
'wallet_exposure_limit': 0.2}}
n_days 61.0
starting_balance 1000
backtesting...
Traceback (most recent call last):
File "D:\bot\psv57\backtest.py", line 219, in
asyncio.run(main())
File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.3568.0_x64__qbz5n2kfra8p0\lib\asyncio\runners.py", line 44, in run
return loop.run_until_complete(main)
File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.3568.0_x64__qbz5n2kfra8p0\lib\asyncio\base_events.py", line 647, in run_until_complete
return future.result()
File "D:\bot\psv57\backtest.py", line 215, in main
plot_wrap(config, data)
File "D:\bot\psv57\backtest.py", line 68, in plot_wrap
fills_long, fills_short, stats = backtest(config, data, do_print=True)
File "D:\bot\psv57\backtest.py", line 47, in backtest
return backtest_neat_grid(
ZeroDivisionError: division by zero

[donewiththedollar]

Hey try a higher start balance in your backtest config. ie. 10000

That should solve the issue :)

[ilaykim]

That solves the issue.

btw I've been debugging this for a good 4 hours now, trying to figure it out :)

[ilaykim]

@donewiththedollar

Quick catch and BAD BAD WORKAROUND the bug

line 42 in backtest.py

elif passivbot_mode == "neat_grid":

    try:
        return backtest_neat_grid(
        data,
        config["starting_balance"],
        config["latency_simulation_ms"],
        config["maker_fee"],
        **xk,)
    except ZeroDivisionError as err:
        print("ZeroDivisionError")
        config["starting_balance"] = config["starting_balance"] * 2
        print("changing starting_balance to", config["starting_balance"])
        try:
            return backtest_neat_grid(
            data,
            config["starting_balance"],
            config["latency_simulation_ms"],
            config["maker_fee"],
            **xk,)
        except ZeroDivisionError as err2:
            print("didnt work!")
            return None