Hey Guys, I am new to this forum and just saw this post. I have been trading almost exclusively CL, and I have been working on a strategy using tradingview for the past 2 months. I am in the process now of developing a bot. But as I am sure some of you know, it is easier said than done. I have it made in tradestation, but I cannot stand tradestation's desktop app. I am woking on integration with the API instead, and it will run on a server in conjunction with tradingview webhooks/alerts.

The strategy runs on 5 minute chart and has very small trades. Avg win after broker and fees is $30 ($0.03 on the CL chart). It wins about 85-90% of the time, however, the stops are a much bigger than the profits.
Per tradingview strategy tester, this wins in a consistent incline on their strategy testing report.
I am still trying to find more ways of reducing the stoploss.
Also, since I am using the 5 minute chart for this, I can't go further back than about 20,000 candles which is about 2 months or so.

It also works well with some input changes on NG, SI, ETH, and RBOB. However, RBOB the stoploss is much smaller and is actually less than the profit target. However, the percentage of wins is much less. Still a winner.


I am willing to share with someone else that wants to share their strategy and exchange ideas. However, in private messages, not public.

I attached a performance report from tradingview here.

Thanks for the post, and in the spirit of trying to understand the power of the post, may I ask, as a NinjaTrader (only) user, if the strategy can be "translated" into NinjaTrader terms, please?

I strongly urge you to reconsider this strategy. These markets are not for scalping 3 ticks in WTI when the bid/ask spread is 2 ticks minimum. And RBOB? The spread is at best 15-20 ticks, its double that overnight. Considering that you backtested on 5 minute bars, I am guessing your strategy assumes it gets filled at the closing or opening price of these bars. Even with a bot, it is very unlikely your strategy will be profitable because you wont be able to replicate your backtested fills.

I am new to doing a lot of this stuff, so I am sure that I may make some rookie mistakes. However, I am very good with programming/coding and math, so I can sit and play with this all day.
The one thing lacking tradingview is a solver that can automatically play with the inputs until it finds a good combination.

Since you are new and you are good at math and programming, consider this: there are 2 things needed to make a model - data and math. If you use the same data and the same math as everyone else, your results will be the same as everyone else and so will your performance. And we already know how the majority performs. Unless you can invent new math, get creative with data. Building your own backtesting tool and working with raw data (not data via some platform) to generate your own, unique model data will be the best investment you make in this process.

For getting data, there is a new startup that I just came across called databento (databento.com). For CME only, they seem to have whatever granularity of data one would want from just end of day OHLC to MBO. Pricing is very reasonable.

Sorry, I'm not a Ninja programmer.

Curious as to whether anyone has been able to code a profitable trading bot specifically for CL or MCL futures? I have been working to accomplish this, and wanted to see if anyone else had a successful bot they were willing to share. Appreciate it!

I've tried different ways of forecasting using ema's of vol over different time periods.