Seeking Community Feedback - How does my trading bot stack up?

[putrid_ask_3791]

Hi everyone,

I’ve been working on a trading bot as a personal project for the last few months, and I’m hoping to get some feedback from this community. My goal is to develop a tool that can autonomously trade assets while adapting to different market conditions. Since my background isn’t in finance or trading, I’m eager to hear thoughts from those with more experience in algorithmic trading. The bot appears to work well in my tests, even when the market is not performing well, it still manages to accumulate more volume for cheap.

Some other interesting features:

· The approach is not deterministic, meaning that double the starting budget does not mean double the performance. Generally, the larger the budget, the better the results.

· Features machine learning optimization for parameters that are optimizable

· There is an option for three trading strategies (conservative, balanced, aggressive)

· Accounts for vendor’s commission and admin fees for accurately reflecting net performance

I’m not here to promote or sell anything—just genuinely looking to learn and see from people that know more than me where my approach stands in terms of results. I’ve also developed a testing environment where anyone can experiment with the bot and upload their own data. If there’s interest, I could consider creating a REST API for broader use. Again, I’m not here to make money or promote anything just looking for some genuine feedback.

Thanks for taking the time to read this, and I look forward to any insights or advice you might have!

 

[Rufus_Leakey]

It's a little less aggressive than I would have hoped for, but at least it have any losing trades.

But seriously, since I couldn't get it to show any trades, perhaps something was wrong with my data.

 

[putrid_ask_3791]

It's a little less aggressive than I would have hoped for, but at least it have any losing trades.
View attachment 337402
But seriously, since I couldn't get it to show any trades, perhaps something was wrong with my data.

Thank you so much for trying it. Nothing wrong with your data, just a silly bug that's now been resolved I would appreciate very much if you tried again and let me know what you think

 

[Rufus_Leakey]

Thank you so much for trying it. Nothing wrong with your data, just a silly bug that's now been resolved I would appreciate very much if you tried again and let me know what you think

Now it shows trades, but the result lost the X axis label.

The "Overall Profit" reported after a run appears to be the equity (value) of the account and not the profit.

The full results data would be more useful if it had columns for the volume held and equity. That could make it easier to calculate a risk-adjusted return.

The ending results for the attached not_random_at_all.csv with commission 0.1% and trading budget 10000 is
Overall Profit (Volume Held * Current Price + Current Budget): 10600.3 for the conservative strategy
Overall Profit (Volume Held * Current Price + Current Budget): 14284.48 for the balanced strategy
Overall Profit (Volume Held * Current Price + Current Budget): 14918.85 for the aggressive strategy

The approximate buy and hold results starting with the entry row for the first trade entry
93,9965.07,9.81,FIRST_BUY
until the last price
66.82
is about 10000 / (9.81 * 1.001) shares * 66.82 last price == 68046.12

I can't judge any risk-adjusted return for the system's strategies, but the buy and hold profit (68046.12 - 10000) would have been 11.8 times the most profitable strategy's profit (14918.85 - 10000).

 

[putrid_ask_3791]

Now it shows trades, but the result lost the X axis label.
View attachment 337408

The "Overall Profit" reported after a run appears to be the equity (value) of the account and not the profit.

The full results data would be more useful if it had columns for the volume held and equity. That could make it easier to calculate a risk-adjusted return.

The ending results for the attached not_random_at_all.csv with commission 0.1% and trading budget 10000 is
Overall Profit (Volume Held * Current Price + Current Budget): 10600.3 for the conservative strategy
Overall Profit (Volume Held * Current Price + Current Budget): 14284.48 for the balanced strategy
Overall Profit (Volume Held * Current Price + Current Budget): 14918.85 for the aggressive strategy

The approximate buy and hold results starting with the entry row for the first trade entry
93,9965.07,9.81,FIRST_BUY
until the last price
66.82
is about 10000 / (9.81 * 1.001) shares * 66.82 last price == 68046.12

I can't judge any risk-adjusted return for the system's strategies, but the buy and hold profit (68046.12 - 10000) would have been 11.8 times the most profitable strategy's profit (14918.85 - 10000).

Thanks, very helpful.
I wonder what do you think of the overall results?
I suppose when you design a system to trade (and assuming no inside info or ownership of a crystal ball ) it's not possible to allocate all your trading budget at the lowest price and sell at the highest.
So I am wondering how it stacks up with other fully autonomous trading systems or bots if you'd like.
Some other details, as I mentioned in the original post, the results are non-deterministic and higher starting budget does yield higher equity at the end, which I think is another benefit.
Having said that I believe there are more improvements that can be made for improving its performance.