throwawaykyok

I agree with you, but if this is the case, isn't it a bit sad that YC funded something like this (or, equivalent, allowed the pivot into it)? Would YC fund a casino?

> Nobody serious trades the candlestick patterns you're referring to in algo trading.

You are correct. My definition of TA was these candlestick patterns, like the ones ("bullish engulfing") shown on the landing page. Mean reversion, momentum, etc are fundamental concepts that have solid mathematical foundations. I have absolutely nothing against these ;)

I have a bit of experience trading derivatives, but little compared to crypto. Everything I said was specific to crypto only. A crucial distinction between crypto and equities is direct market access. Everyone in crypto gets DMA and API access with a few clicks only. Most of the low-hanging arb fruit is thus already taken or getting harder to take and moving to higher frequencies. There are still low-frequency arb opportunities in crypto, but for many of them the edge is in efficient/safe fiat currency movement across country orders, i.e. having the right citizenships and bank accounts and clearances. In many countries it's not easy to open bank accounts for crypto trading these days.

People often think there are arb opportunities in crypto when there are none because the price already includes the inefficiencies, latencies, and difficulties of moving fiat across country borders or taking money out.

Also, the price often includes the risk of the exchange being hacked or running away with your money, which has happened a lot recently. So when people look at arb opportunities they often don't take into account that these risks must be reflected in the exchange prices. Sketchy exchange prices are lower due to the risk factor of having balances there. That's not the case for regulated financial markets.

Yes, that's a valid use case for this tool. Exchanges typically don't have this functionality. My opinion based on my personal experience is that it is rather unlikely to build profitable algorithms based purely on "technical analysis patterns": (1) Most of these patterns are pseudoscience, you can google for more info, only a few have even a bit of science behind them (2) There isn't enough data to train a model on minutely or hourly data. The crypto markets change extremely quickly (data distribution shift) so that most of the data you would be using to optimize model parameters is outdated and leads to false conclusions.

To make money off less frequent trades I believe (again, just my opinion) that your decisions must be based on news/insight/insider info, not charting patterns. In other words, more fundamentals. And that's much harder to automate.

Of course, you will find people that tell you the opposite and that they make money. My response to this would be: Of course there are those that make money due to simple laws of probability. With a lot of people trading there will be some winners. But it is mostly due to luck rather than skill. And the fewer trades your make (lower frequency) the harder it is to assign any kind of significance numbers to the results.

What you have built looks great and I am sure a lot of work went into it. However, as someone with a bit of trading experience (running a profitable custom-built trading system myself) who has friends that lost money I would like to post a few words of warning to anyone who believes they can build profitable trading strategies on top of someone else's platform: It's extremely unlikely. While you may hear stories of people making money, most of them are pure luck (I recommend the book Fooled By Randomness). To make a stable incoming with trading you must have a consistent "edge", a competitive advantage that other traders don't. Possibilities here include 1. data, that could be be cleaner, more fine grained (L2/L3 book data, better reconstructed, etc) 2. infrastructure. This includes highly latency-optimized server placement, custom API integrations, fault tolerance, dealing with api issues, etc 3. "smarter" trading strategies - What many companies are selling you is that you can make quick $$$ by coming up with some secret trading strategy (3). That's how they make money off you. These charting patterns are pseudoscience. 99% of all edge in crypto trading is in #1 and #2 - infrastructure and data, with relatively simple well-known strategy algorithms. By "outsourcing" this to a platform you are giving up your edge and set yourself up to lose money in the long run.

Also, ask yourself why someone would offer a platform to build profitable strategies instead of simply trading themselves based on their competitive advantage. The answer almost always is: Because they failed to trade profitably and pivoted to selling their (unprofitable) infra.

I don't want to put down what you have built. I know firsthand how hard it is to build some of this infrastructure. I would just like to warn people to not easily trust trading infra providers before they lose money.