Launch HN: Mudrex (YC W19) – Automated Trading Without Code

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.

To everyone pointing out that you need an edge to trade successfully, you can’t beat HFT guys with this etc etc...this is a tool for dumb money. Just look at casinos, they make billions from dumb money. Just because you can’t beat hft traders with this doesn’t mean there isn’t demand for a tool like this. People will buy this so they are making it.

Are you aware there is a massive (market leading, in fact) trading platform called Murex? The name is rather close isn't it? Is it intentional?

Edit: I mean, if I started making say, an operating system and called my company one letter different from another company (mickrosoft? rehhat?)-- who are in the same market and have an enormous reputatation -- I'm sort of asking for trouble, aren't i?

Are there really people out there who can create profitable automated trading strategies with all that involves, but can’t code?

Presumably they can’t test either and arent too hot on maths?

Who are these people?!

> We dream of reducing the gap around financial opportunities between the wealthiest and the rest.

I know it's one of those lines that whoever is in the CEO role says really righteously over and over again whenever an investor pushes back on it.

But is trading cryptocurrency really going to do that? Like normal people actually lost their life savings trading that. And it's rich people, like actual Silicon Valley supervillains of sorts, who generated a very large amount of these imaginary assets they're selling to normal people. And then, I'm not sure transitioning to stocks is any better, because anyone buying today is pretty much guaranteed to be buying at the top of the market--i.e., if you're making a market, you're making one for suckers.

Are you just reducing the gap around financial opportunities to lose a lot of money instead of a little?

Sorry this is a bit off topic, but how do crypto traders manage their taxes? I am under the impression that crypto trading doesn't have any of the niceties that traditional securities trading has when it comes to taxes, e.g. only paying capital gains when you remove money from your trading account (maybe that is also incorrect, I have no experience with investing).

Do you have any plans to support or integrate sports trading?

I've been creating trading bots for a while now and would have preferred if you targeted your communication at someone like me rather than at people new to the market. It would have seemed more honest. As it stands you don't address any of the potential questions I might have had, such as:

Why the focus on crypto when it's illiquid, unregulated (people lose money from folding brokers regularly), hard to secure etc. vs forex which is highly liquid (very little gapping/get the price you want, when you want), very well regulated with many highly professional international brokers offering APIs you could work with right now. It feels disingenuous to target crypto only, even if initially, like you're cashing in on the (currently subsided) wave. I've talked to 100s of traders too and I don't recognise the problems you've raised.

Why mention 1000s of currency pairs when... there aren't 1000s? In forex there are dozens, in crypto - well OK any amount a broker puts together, but understand that the more rare or minor the pair, the less liquid and harder to trade by algo. And you don't want to algo trade correlated pairs, that's just taking more risk unintentionally. So in practice you don't want to be trading 1000s, that's unrealistic and misleading.

Where's your education section - this really isn't as easy as you make out, and you offering MA Cross as a sample shows how much this is targeted at the beginner level. Which is where people lose money. You need to be the resource for algo trading education. Better than the ones offered elsewhere.

Saying you're democratising something that is pretty much already democratised is not a great start, IMO. Algo trading in forex is not an immature market. Look at the Metatrader website or Forex Forums and you'll find dozens of free bots to get you started with the code and backtesting.

Going forward I'd like to see you educate your users better, talk more about the realities of algo trading and offer more up front.

You've probably built something really cool, but I can't get past these initial red flags. You should show us the non-coding interface and talk up that side. I've used similar in the past and it can be a quick way to build an algo, but you've got to have guidance to make anything profitable - I've been through 1000s of ideas but the markets can offer multiples of those back in edge cases where they fail. You should list your indicators, talk about what risk management and order management features you've come up with, show traders how you've solved many of the tricky problems.

There's huge room for a profitable business in this field, particularly in creating a platform that can displace MT4 with a multi-core, secure backtesting environment using Python or similar. Maybe even incorporating ML in a non-code fashion. But the non-code parts would be the icing on the cake. Most algo devs have to get their hands dirty at some point.

Otherwise I'm just worried you're going to mislead a lot of people.

That's my 2 cents, hope it helps with perspective for you or anyone coming after you in this space.

How much data are you extracting or planning to extract from profitable trading strategies? This seems like a great way to crowdsource market intelligence.

To quote another analogy below - It's the modern version of selling shovels, except the shovels have GPS and mass spectrometers installed, so you can later pull in from behind with excavators and dump trucks.

The modern version of selling shovels I suppose. Interesting idea.

Congratulations on the launch. Beautiful UI! In terms of product market fit I can tell you that I've been kicking around building systems to automate my investment decisions, Ray-Dalio-style. I personally adhere to the church of macroeconomic investing, so what I was aiming to build was a setup that would buy or sell positions when various leading indicators went one way or another. It'd be really cool if you hooked into Google Sheets. My plan was to run nightly Google Scripts that dump the data into Google Sheets and get alerted when thresholds have been crossed or trends are emerging. Or, you could pull this data into your platform and enable users to make decisions based off of it. Pretty much everything I need is available from FRED, but if you got access to proprietary data like the Consumer Confidence Index that would be a big value add for me.

This next comment might make me some enemies, but seeing as I seem to be close to your market I figured you'll appreciate the feedback. Crypto seems like a dying fad to me. More importantly, I'm weary of doing business with companies engaged in crypto. I loved Robinhood up until they jumped on the crypto wagon. I still use them, but I fear that something crypto-related will blow up, and my hard-earned money will suddenly be tied up in an insolvent startup. Irrational fear, I know, but it's there. The real opportunity here in my eyes is to go after Robinhood, and build a similarly user-friendly experience for slightly-more sophisticated users.

if you ask me, is this a good idea? I will tell you it's a terrible idea. If you ask me, could this startup make money? sure, there are lots of people out there who would pay for this.

Looks a lot like kryll.io. The concept is the same, the UI seems very similar. I understand that it's a YC project, so it does already have a good chance to work just from the networking effect.

But other than that, how do you differenciate from kryll, since frankly, they are already pretty good at what they do (the shared strategies, backtracking and over all the community is very nice) ?

Does your strategy creator allow for multiple time frames to be considered simultaneously?

What do normal algo trading models look like? Are they all super proprietary so it's hard to talk about, or are there useful and representative models available for learners to dig into?

This is great execution and UX. Congrats for launching!

As someone who has done professional algorithmic trading before, I would strongly advise most users to deploy a small amount of capital to these strategies and treat it as a fun game, especially in the beginning. The most important thing in automatic trading strategies is risk management - I am not sure what intelligence does Mudrex provide around that, but even if you're right 70% of the time (very high for automatic trading), the rest of 30% trades gone wrong can sink you quickly if there are no intelligent ways to manage those trades.

The returns and losses also highly depend on volatility of instruments (volatility refers to the average uncertainty in returns as measured by Statistical Standard Deviation). Cryptos are highly volatile. Cryptos are fun to start with, but you must allocate money keeping that into account and you'll be fine. My recommendation is at least 15-20 times less for Bitcoin than what you would allocate to S&P 500, based on their volatility/risk factor alone (15% vs 75% annualized volatility). And even less for other coins (See project below for more guidance)

As an aside, I have been working on my own project for many months to bring trend detection, risk management and intelligent diversification for all DIY Crypto and Stock investors [0][1]. Feel free to check it out if you invest in stocks or crypto at any level. It's in Alpha and will be shared more widely in the coming month or so.

I've thought about bringing automatic trading to my venture, but after a lot of my own time spent creating my own strategies, I can't sincerely recommend it to others without an extremely high level of skill. The hedge fund I worked with had people watching these automatic strategies all the time with lots of alerts, triggers and Plan B management, even if they were "automatic". Who is going to watch the automatic strategies here?

[0]: https://stockquanta.com

[1]: https://coinquanta.com

I’m genuinely curious about who the target customers are for this platform. In terms of simple strategies (see signals A and B, fire order, wait for signal C, exit market, profit), there’s just no chance of a startup beating the established HFT players on execution speed. In terms of sophisticated model generation, training, optimization, etc., I don’t see being able to compete with them either. Their data and computational resources are just too good. You interviewed 100s of traders; I wonder how many truly successful ones you talked to. Did you give more weight to their input than the less successful ones?

I’m certainly biased here, as I’ve spent many years building infrastructure for large HFT firms. If nothing else, it’s made it crystal clear that there’s smart money in the market and dumb money in the market. On my own, I’m absolutely the dumb money when I compare my resources to infrastructures costing $300M or more and hundreds of the brightest people around. Unsurprisingly, I keep my money in mutual funds.

I wish you the best of luck, though. I hope you prove me wrong.

Fascinating, we were building something similar here at KloudTrader a long time ago.

https://medium.com/@kloudtrader/private-beta-and-kloudtrader...

Let's talk?

https://kloudtrader.com

In your beta thus far, what’s the mean and median returns for user driven strategies?

Why isn't this just a Quantopian algorithm hooked up to the actual Robinhood API when it comes time to buy/sell/deposit more funds?

The challenges you listed are spot on. Congrats on launching! Can’t wait to check it out.

Curious in terms of unique users, what % lose/make money over time? How long do the losers stay on platform before exiting?

Minimal look is quite slick. I would like to to test the cloud developer interface using a demo account to provide feedback ;)

What is required to integrate additional exchanges? Would you eventually want more?

Great team! Excited to see where this goes.

Aside from the few reputable exchanges, most of the volume in crypto is fake.[0] How do you account for that?

[0] https://www.blockchaintransparency.org/

Cool! What sort of liabilities do you face with this?

For example, let’s say my strategy has some sort of logic to exit all positions if X happens.

X happens, but your platform goes down, so my positions stay open and my account gets wiped out/suffers large losses. Am I SOL or do you assume some responsibility for that?

My usual response to launching a startup is congrats but this seems a really cynical value extraction tool that actually probably destroys more value than it creates as it charges a vig along the way. For shame YC

Many people are criticizing mere existence of this project. I would like to however offer my support to the creators of this tool. I do fully admit number of people who will use Mudrex in next 12 months and will make passive income of thousands of dollars per month will be 0. Almost 99% of the users are going to lose their money if they do live trading with this tool. And yet, I think this tool makes the world a better place.

Blaming Mudrex for people losing money is like blaming government for building road towards Casinos. Mudrex itself is not like a Casino, that would be Crypto exchanges and price movements.

Secondly, there are many many positives of tools like this. You can buy some basic books on algorithmic trading and implement some algos and gain a lot of experience in how algo trading works, what different models exist, if the models fail why do they fail and if no one makes money on algo trading why that is the case. There is a stiff learning curve that in my opinion is very valuable in itself. You will not get any new information if you lose money on slot machine but if you have spent enough time putting efforts and learning algo trading you might end up learning a lot. That knowledge itself could be very useful.

I do not expect people to invest their lifetime savings in Mudrex. Most people might spend only few $100 on this and few among them will be serious and will gain a lot of knowledge.

Kudos to Rohi, Edul, Prince, Alankar and Snehil. I wish you good fortune in battles to come. I did play with the tool extensively and I liked the design, how fast it is. I have following suggestions:

1. Whenever I click on a Backtest I should be able to fork the very specific strategy that was used for it and not the current strategy. In short, support strategy versioning. Most often I design a strategy, run a backtest, change the strategy run a backtest and I have like 10 backtests running for different variants of the strategy. Currently there is no good way to go back to the specific varient that a backtest was running.

2. Let me download some file of the strategy. I should be able to send it to my more knowledgeable friend and get feedback.

Thanks for sharing. My advice is shut it down. You are making the world a worse place. Just because you can give people the ability to gamble more does not mean you should.