Every AI trading bot I’ve ever tried lost money, except for the ones I programmed myself. That was true a couple years ago and it’s true now. What’s the difference worth writing about? I’ve opened up access that everyone can take advantage of. Finally, something that works.
Most AI Trading Bots Fail
It’s a tall order to automate trading for profit. Trading is hard. Coding trading bots is harder. Coding trading bots that actually turn a profit is extremely hard, yet here we are. Although my first few bots were terrible, these were simple rule-based engines that didn’t do much, save frustrate me in the late hours of the night. The game-changer was finally getting good at trading and using my artificial intelligence skills in my efforts. An AI trading bot is something approximating a holy grail of trading. I didn’t used to have much hope for that sort of thing. I knew AI, sure. The problem was that every post I read said it was futile to attempt. As I would soon come to find out, those posts were very (And possibly deliberately) misleading. By the way certain people wrote their whitepapers and Github README.md files, I got the impression that the good code was being hidden. Not only that, some coders even seem to go out of their way to discourage other people from getting on the right track algorithmic trading.
My first glimmer of hope came from a YouTube video. It was a Numberphile interview with a man named James Simons. I learned that Simons is a legend in his on rite. Would it surprise you to know that one of the biggest and most successful funds in the world depends completely on AI trading bots? It’s true. Mr Simon’s Renaissance Technologies is apparently a legend in quantitative finance. Renaissance, launched in the 1980s a success – remarkable given the technology of the time. As I watched the lengthy interview, my eyes glossed over. Simons was a mathematician; I love math. His company relies solely on algorithmic trading; I love to code algorithms. That was the spark I needed, hence my progression.
From AI To TradingView Strategies & Beyond
Fast-forward a few months and I had managed to build out a working AI, cryptocurrency prediction platform called ANNA. At one point I was selling access to cryptocurrency AI forecasts. The proceeds from that wasn’t the goal as I mainly wanted to help people – showing them the possibilities. ANNA is still a work-in-progress but that’s besides the point of Pinebot. Pinebot is very different yet produces nearly equivalent results. Imagine if the two were could benefit from each-other.
I learned a lot from my time coding the AI platform. Eventually I began to experiment with TradingView strategies. That’s where Pinebot got started. My first experience with TradingView strategies was a bad one. It wasn’t the fault of TradingView. Rather, every user-created strategy I bought claimed the world and delivered nothing. I had to start coding my own strategies to realize why. Many (Not all) Pine script strategy developers probably don’t realize or don’t want to acknowledge that their code leads to misleading backtests. Consequently, this was a letdown. Not to be completely dismayed however, I soon realized that success is possible.
My first good strategy was meant to be a supplement to my other AI trading efforts. The problem was, my AI code had no way of talking to TradingView. The path was clear. I needed to put my other AI efforts on hold. With a Pine script trading platform, every other thing I coded could benefit from a kickass backtesting engine. Thus, Pinebot was born. I caught a lot of flack for pausing the AI project but like I said, the path was and is clear.
Ego & Risk
Though many strategies are duds, some are really good. Coders on TradingView who are willing to be honest with themselves and the people they serve will quickly realize real success. Not mere novelty but the profitable automation we seek. This is the reward for seeking honest backtesting. Of course, the TradingView backtester isn’t perfect. It doesn’t need to be though as it gets the job done. That being said, one does have to keep one’s ego in check at all times. “Maybe I’ll just let this one repaint a little” is fine if you’re prepared to compensate for it. What’s important there is that you have to know what you’re doing to even begin to levy those kinds of risks in your code, in your favor. That’s a fine term there as well: Risk – because risk management is the cornerstone of every good strategy. Following suit, this is exemplified in every strategy I put out with no exception.
Us vs Them
So how does this all compute an algorithmic trading bot that actually works? Pinebot can make use of any TradingView/Pine script strategy. It uses their long/short and open/close calls to make trades for you via your exchange account, using your API keys. Recall every useless alert script and there’s no comparison. Also, I took my best strategies and posted them up for sale here. Along with it also, a Pine script marketplace awaits. For the first time ever, any Pine coder can instantly setup their own shop and start using their strategies as bots within minutes.
So far, my testing has gone extremely well. Likewise, every trade placed has been at or near in line with the expectations of the backtester. With Pinebot, I’m actually turning a profit. Contrast that with the disarray of bots I almost all lost money on: Gunbot, Autoview, Profit Trailer, AITrader, Cheetah, Tortuga, CryptoHopper, HaasBot, and others. The list goes on. I figured someone had to have gotten it right, right? Wrong, at least publicly. If you want something done right and all. What’s unique here is, for one, that I’m making public something that works. I rather suspect that’s a rarity. Also, I’m giving others the ability to help it improve, grow, and they along with it.
Pinebot VS Other Bots
Works Without Software Download
Simple Order-Book Trailer
Pinebot Rises Above
While I am proud, the purpose of this article isn’t to gloat. On the contrary. Perhaps mitigating factors in some of those other bots contributed to my experience. Who knows? My old jaded self would have thought “Great, a brand new elaborate waste of money!” Above all, Pinebot is not that. To be clear though – people who sell trading bots they know don’t work good are certainly despicable (Discounting open-source projects, of course, respectfully). Perhaps that or they’re massaging their own egos.
This platform isn’t perfect (Yet) but it’s good. I sometimes have to temporarily take the bot down for upgrades and stability maintenance. However, that will happen less and less going forward. Pinebot is as strong as the strategies coders post here (Mine included). There are plenty of noteworthy strategies that work well. In the past few days, our testing has quadrupled the balance our Bitmex account. That’s a remarkable achievement. If you’ve been burned by other bots (As have I) try us out. It’s free to try, after all. Even more…
The maths prove Pinebot. Finally, an algorithmic, AI trading bot platform that actually works.