Hacker News new | past | comments | ask | show | jobs | submit login
Launch HN: Mudrex (YC W19) – Automated Trading Without Code
157 points by rohitgoyal on Mar 9, 2019 | hide | past | web | favorite | 114 comments
Hello, HN!

We're Rohit, Edul, Prince, Alankar and Snehil. We’re building Mudrex (https://mudrex.com).

Mudrex helps traders automate their trading without having to code and spend a lot of time and money building infrastructure. Though we have started with cryptocurrency trading first, our goal is to help anyone who wants to get into automated trading or investing across any asset class.

We had been trading cryptocurrencies for some time and were not able to track trading opportunities across 1000s of currency pairs for 24 hours a day and 7 days a week. We thought of automating our trades, but faced challenges like being unable to access historical data, building a testing framework to test strategies on historical or live data, and maintaining connections and orders on multiple exchanges. Also, since it was all coded we were not able to quickly iterate on trading ideas, since for simplest changes we had to dig into code and deploy new strategies to retest them. That also made it hard to have a log of all the trades, strategy changes, back-tests, papers to compare things at one place.

After talking to hundreds of traders, we found out almost all active traders were facing the same problems, so we decided to build Mudrex. At Mudrex traders can quickly build any kind of trading strategies using 150+ indicators, 50+ candlestick patterns, price or volume action. They can test their trading strategies on historical and live data to optimise and automate their trading by connecting their order using exchange API keys.

Cryptocurrencies are just a start for us and we have been growing very quickly since the launch and getting feedback from more and more people every day. There are millions of traders out there, trading manually across different asset classes just because they don't have capital and resources to build an automated trading infrastructure from scratch. We want to solve all of their problems so they can focus on building trading strategies instead of spending too much time on infrastructure issues. One of our users defined our goal as "democratising algorithmic trading and investing" and we are working hard every day to get closer to that.

For the next 2 months the platform is free to use but we will charge a $0-$100 monthly subscription fee depending on usage with a free tier available.

Next on our roadmap are things like a code editor (where traders can build custom indicators using minimum coding), a marketplace between traders and investors, and integrating equities and forex.

We dream of reducing the gap around financial opportunities between the wealthiest and the rest. We just got started and are going to face many technical/design/operational challenges, but with the help of our users' feedback and support we feel we can achieve it. Looking forward to hear feedback from the HN community to keep going in the right direction!

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.

> 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.

This is correct. It's so obvious but most people simply don't see it for what it is.

The interesting thing about these platforms is that they can actually record what OTHER people are doing, see what the successful people do, and possibly just copy them.

I'm really glad these points are being pushed in this thread--it is almost impossible to tune a portfolio for performance without knowing something the market doesn't (data, computing power, information). That's why I only ever tune my portfolio for the level of risk that I can tolerate. Because that's pretty much the only thing you _can_ tune for.

What you are saying is right specially for HFT and also because no platform have been able to achieve that till now. We decided to build this only after talking to 100s of traders personally. We definitely can’t help people develop profitable strategies but if someone has a strategy, she wouldn’t be stuck because of resources/capital limitation. Having said that we are going to keep building the state of the art infrastructure to support more efficient trading as well. Thanks for taking so much time to provide feedback.

I don’t think this really addresses his criticisms of your startup...

Threads are going off the rails a bit.

You built a fine tool for others to use. With a free tier. Nice work...

Thanks for your support. :)

that has nothing to do with what hes saying

In response to your second point - it's largely because it takes quite a bit of capital to build wealth via trading. Much more capital than is needed starting a SaaS company.

It's hard to win picking individual stocks, and yes you should not only hold a few stocks at a time with all of your portfolio.

But you don't have to bet your entire portfolio. You can only trade a portion of it hedging your risk. In addition, while it may be more of a fools game to go big on one or two stocks, you can look at your risk profile and align your long term strategy with different sectors and tax strategies.

I am no means a trading expert, but to discourage people from investing just because certain types of trading is more akin to playing the lottery - is just as irresponsible. People should be encouraged to be more engaged with their money and at the very least follow how their index fund is doing, how it's allocated, and learn why it is allocated the way it is.

I think GP is cautioning against high frequency algorithmic trading, not investing in general.

Totally naive question: what if you aren't interested in HFT but just want to take some of the manual labor out of making less frequent trades? Is that any less reliable than sitting at your computer watching the order books or do the exchanges already have all the tools you would need for that?

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.

It doesn't sound like you've thought this through.

Technical analysis isn't typically based on patterns. Banks trade using TA (see Lehman Bros Forex trading handbook - available online) - it's not strange or unusual to trade a mean reversion strategy at HFT or retail trading levels. Nor a momentum strategy, or even trend following. Nobody serious trades the candlestick patterns you're referring to in algo trading.

There is enough data to train all forex pairs - tick data going back decades, for free, via Dukascopy and other sources.

You're spreading FUD, and while there is definitely a need for people to be cautious, your warnings are not based in reality.

Better to warn them about the 70-90% losing rate of all retail traders (reported officially by all UK brokers), or the perils of overleveraging, or the risk of gapping/illiquidity in crypto... I could go on. But not what you mention - that's a strawman.

> 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 ;)

OK, OK, misunderstood your point - apologies. Yeah if they're selling those as potential profitable avenues... red flags are up. That's not selling shovels to miners, that's selling spoons.

That's a screenshot you are referring to. It just shows how to connect different blocks. Like see the first gif that shows the kind of blocks we have with an example of indicator block (RSI/MACD etc). In no way there is a sign that it will make you profit. In fact it encourage you to test your strategies on past data and live data. Will try to make it clearer.

I disagree with the notion that your edge has to rely mainly on the infrastructure or insider knowledge. In today's markets you often have the case that very wealthy and high volume market participants have to make ends meet with rather low liquidity markets. This leaves more than enough room for small fish to arbitrate since the price allocation is flawed.

Out of curiosity, do you also have experience trading equities/derivatives? I largely agree with you (especially the words of caution and about buying other peoples' trading infra), though I'd push back a bit on your assessment of automated trading at lower frequencies than HFT. I'm in strong agreement with you about the lack of utility for what is typically called "charting" or "technical analysis", but I don't think it would be fair to characterize all of stat arb at less than HFT timescales this way.

But my experience is entirely outside cryptocurrencies, so take that for what you will. I'm just curious to hear if you're also applying this perspective outside cryptocurrency trading. If anything I'd expect it to be far easier to profitably trade cryptocurrencies these days since the space is still so inefficient.

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.

Just curious how much money you made in total and you’re sharpe ratio? What about win rate and how man trades you made? We could use that to determine if you have an edge (not saying you don’t).

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.

>Why the focus on crypto...

We started with cryptocurrency because of our personal experience in cryptocurrency trading as we had to stay awake sometimes to close our positions, we missed trading opportunities with manual trading. We felt the infrastructure is much more limited for cryptocurrency trading since it is new and majority of the trading is done by individual vs institutions compared to other asset classes. Also the market inefficiencies are greater in cryptocurrency trading and individual traders or small shops have higher chances of building profitable strategies where they don’t have to compete with HTF at the moment.

>Why mention 1000s of currency pairs...

You are right. 90% of the trading happens on just top 20 coins but since our users requested other pairs we added them.

>Where's your education section...

Point well taken. We are already producing content on the education side. https://medium.com/mudrex/tagged/learning We are planning to launch a more formal education section for new traders.

>Saying you're democratising something...

Right now traders are dependent on coding skills and capital to set up infrastructure to automate trading. For example a person with a balance of $1000 can’t get into it.

>You've probably built something really cool...

Your concerns are valid. We are doing webinars, writing blogs and trying our best on this side and will do things as time progress. https://mudrex.com/indicators https://medium.com/mudrex/managing-risk-while-trading-4c1455...

>There's huge room for a profitable business...

This is exactly what we are going to build. We are working on a code editor where people would be able to build their custom logic while keeping the simplicity of the platform same for everything else. On a high level our objective will remain “building tools traders need in their automated trading”.

Thanks for your time in providing the insightful feedback.

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

You are absolutely right. And for that reason we have worked hard to build risk management tools. You can read more about it here on our blog as well. https://medium.com/mudrex/managing-risk-while-trading-4c1455...

If you have any other suggestions as well please let us know as we will keeping building on it.

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.

We are currently focused on strategies that typically have exposure from few minutes - few days. These don't come under the HFT set up.

Hence our target customer are people who use Technical analysis and price action.

Thanks for you time and providing the feedback.

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?

Your comment was dead, so I vouched for it. I'm not sure if your point is valid or not, but I'm curious what the founders have to say about this.

Kind of curious also how this works. If they’re in YC surely someone must have noticed? Murex is in use in all the banks I’ve worked at in the last 5 years, it’s rather well known?

Didn't know about this. Thanks for pointing it out. Mudrex has nothing to do with it. Mudrex came from the Sanskrit word call "Mudra" which means currency. I hope this answers your concern.

I would be carful; if you think sap people are billed a lot per day you should see how much we’re invoiced for murex consultants.

Not the sort of companies who play nice on these sorts of issues.

Just my 0.02

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?

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.

All the IP of the strategies remains with the user who built the strategy. We don’t extract any data from those strategies.

We do have a marketplace where traders can publish their strategies publicly for other investors to test and invest in those strategies.

Thanks for checking us out.

Buy like how do you guarantee that aside from saying it?

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?!

There are a lot of people who do research using Trading View and optimise algorithms on excel. But they don't know to code and build the whole infrastructure. Thanks for bringing this up.

Not to be a sceptic but you can’t do this ... you can’t optimise algorithms in Excel. I mean you can; you can do whatever you want but what needs to be done is a lot harder than this.

I think you underestimate the mathematical power Excel has under its hood. (personally not an Excel fan)

As somebody who has written software for traders, I think this is an excellent question.

If somebody really has some sort of advantage such that they can be consistently profitable, then at worst they could trade manually for a bit and then hire somebody to code.

Like others, I suspect the real appeal of a platform like this is similar to the value of a racetrack. It's selling hope and optimism to people who think they are smarter than everybody else.

I don't think its a competition between the wealthiest and the rest. Its really competition between a lot of smart people. Some of them might be really good at math.

If you have phd in math I am sure you can figure out how to use python or c, and that is who you are going up against people who have phds from harvard, stanford, Princeton in math, physics, engineering etc. These are the people writing the quant algos going up against yours.

Most hedge fund managers can't beat index funds.

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

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

Spot on, it's a great idea though, one I've heard about from several people years ago. I wonder what the fail-safes are, though. With bot and HFT, I wonder how much can be gamed against them if something like this catches on.

I've also heard of programmers linking trades with bots scanning twitter for possible cryptocurrency trend movements. Smart stuff, but too risky for someone like myself.

We see it as democratisation of financial instruments. Would be great to hear your feedback.

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.

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?

Well yes, I guess I should add that I don’t mean to say this is a “good” idea. I mean it’s a good idea in the sense that people will probably buy it, but bad in the sense that, like gambling, people probably shouldn’t use it.

Of course the people making it don’t want to believe it’s bad and I’m sure can cite statistics and use cases and whatever that aren’t to exploit dumb money. But it’s pretty clear that’s what it is.

While I agree with what you and others are saying, I just want to point out on the generous interpretation, there’s a good chance the people at YC aren’t aware of this. Possible (but less likely) that the companies founders aren’t aware of this also. There’s a lot of naive tech industry people who dive into financial markets thinking they’re like other areas that are waiting for a startup to ‘disrupt’ them - they’re not.

Good point. But maybe being in HN there is high demand also for not inventing something that knowingly takes money from dumb people.

Yes I agree with you 100%. I didn’t mean to condone their actions, just to point out that this is clearly not designed to beat hft, it’s designed to make a profit off of dumb money, and the more interesting debate is whether that’s morally wrong or not. Debating about strategy and hft is a distraction from the real debate which is: should YC fund a company like this?

I don’t doubt that this company can make money, and probably a lot of it. But should they?

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.

What we are trying to build is asset independent. We started with cryptocurrencies because it was easiest to integrate with. I understand your mistrust in cryptocurrency related projects. We don’t hold user’s tokens in our wallet. We are figuring out the best ways to add equities and forex as well.

Right now we are providing all the data feeds ourselves. If there is more demand from the users for manual entry of data, we might support it later.

Thanks a lot for your detailed feedback.

Hi, current macro vol trader & dev here - send me a message here and I'd like to here about your needs and see if my personal setup works for you.

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.

Thanks for your feedback. Would love to hear more on why you think it is a terrible idea.

why do people need to run trading bots ? what's the point? you are not going to win against professional trading bots. And you wont do any better than buy and hold investing anyway, after taxes and fees.

A good startup should add real value to the world. Your startup enables a completely pointless activity.

The platform is for traders who knows what they are doing. We have professional traders building their trading strategies on the platform. We are working on marketplace where non-professionals would be able to invest in professional trader's strategies for profit share.

> The platform is for traders who knows what they are doing.

No, the platform is for people who think they know what they're doing. Which it's why it's a bad idea. It's a very elaborate and expensive way for people to learn that the problem is much more complicated than they initially thought.

>We are working on marketplace where non-professionals would be able to invest in professional trader's strategies for profit share

Sorry, but that makes no sense. If a professional trader created a profitable strategy, why would she share it with non-professionals and "profit share"?

Further, if a number of people started to use a profitable strategy, then it wouldn't be profitable for very long.

sure, if you are targeting professional traders, it's reasonable, but professional traders need to build their own infrastructure to be professional. in trading, infrastructure is the edge. there is no edge in using someone else's trading infrastructure.

I'd invite you to try out the platform then. We aren't offering get-rich-quick schemes. We are just providing the tools to create your own. :)

you are providing the tools to do something no sane person should ever do. I agree you can probably find customers who will do it and will pay you for it, but it's a nonsensical activity like buying lottery tickets.

This isn't targeted at competing with billion dollar HFT firms. A trading bot can simply ping you an alert when criteria are met for you to take one trade a month/year. Or it can be more involved. But it's not instant suicide if you try. Tough, but not impossible. It can be as macro as you want, is my point.

To address your analogy, the odds of a trade going in your favour are closer to 1 in 2 than 1 in 14 million. Of course fees etc. apply, so you have to cut losers short and let winners run, but it's not comparable to a lottery. Or a casino before you go there.

It's very much comparable to a casino or a lottery. In all cases it's a losing proposition for most naive participants, but there's enough gimcrackery that naive participants can think they'll be the one to beat the odds.

The one case in which it arguably wouldn't be the same is if they were playing a positive-sum game, as with long-term equity investment. But since they're focused on hours-to-days trading, it's a negative-sum game.

> it's a negative-sum game

This is a misconception often touted by those who think they know something.

The other side of a trade doesn't necessarily win/lose if they needed to exchange $xxxm to open a factory on another continent. Not all market activity is speculation/investment. Much of it, some say the majority, is random business activity. This skews the game theory analogy enough to make it invalid. There is enough inefficiency, noise and long term trending for a non-naive participant to profit from speculation. For now.

Kudos on the word gimcrackery.

Hi. I have written financial trading software for market makers and have worked for more than one commercial bank, so I'm reasonably familiar with the topic.

I did not say that all trading activity is negative-sum. I am specifically speaking about this, which is targeted at short-term trading in cryptocurrencies. Nobody is using this to open their first European facility. This is for rubes to try to outsmart people.

Whether I'm buying lottery tickets, treasury bonds or using trading bots, I take risk-reward ratio into consideration.

Others have already found a niche in automating trades with bots. Mudrex has found another niche within it by letting people easily create those bots.

In any case, Mudrex allows you to backtest and run paper trades on strategies you or others have created just to help you find a suitable risk-reward ratio you are comfortable with.

Like I said before, I invite you to try the platform. Set the fees and slippage and test your sanity for yourself.

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


Let's talk?


Wow! Would be great to learn from you. I will send you an email.

> 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?

I would criticize the line more directly, in that it's typical silicon valley magical thinking. "The rest" means the slightly-less-wealthy, because these "financial opportunities" are only useful to people with large sums of spare cash. The line reads like a populist appeal, but working class people aren't sitting on HN thinking "finally, improved investment opportunity!"

It's par for the course so I don't mean to call out this person directly, but I'm hitting that point where I don't feel optimistic by default with these claims.

We just started with cryptocurrency trading and plan to add equities and forex soon. What you are saying is right, investing blindly without research make people lose money. Having a disciplined hypothesis driven strategy prevents you from taking irrational trading decisions. Thanks for taking time to provide feedback.

It's probably too late to follow up on this, but dude c'mon, that's not what I'm saying.

> 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.

I am assuming you are bearish and overall negative given the sentiment of your comment. On Christmas eve when the market was at its lows, would you have guessed we would rally and explode up the first two months of 2019. It’s very easy to perpetually be a bear and shout that the market is overvalued. Eventually you will be right and a correction or crash will happen.

I’ve be in the market over 15 years, and can without a doubt say it is the best wealth expanding tool rich or poor US citizens have.

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.

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

Mean: 2.1% Median: 1.7%

Average run time of strategy is 2 weeks in beta.

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) ?

I have checked out Kryll. It's a great project.

We are focused on providing the access of automated trading/investing to everyone.

That's a neutral comment that doesn't commit yourself to anything and doesn't answer my question. Your project assumes your users commit time and money to it, so please take a few minutes to do the same.

I want to save time, so I use kryll for automation. I'm open minded and ok to check out your project, I just don't have time to try everything out. When something works, and works well, I need good reasons to check alternatives.

So concretly, what will I find in your product that is worth the trouble to try a whole new plateform ?

1) Way more variety of indicators (150+) 2) You can create much more complex strategies on Mudrex 3) We support futures trading with Bitmex

So Mudrex allows for more complexity than kryll.

It's true that kryll is very simple to use. If I reach the limit of their tooling, I'll give your plateform a try.

Thank you.

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?

At Mudrex users create their own trading strategies without writing code and connect their exchange keys to trade live.

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).

dislaimer: I'm not a tax lawyer. This is not advice. etc...

In general, trading crypto is taxed like trading stocks or any other commodity. So, when you trade some coins for something else, at that moment you owe tax on the difference in value between when you received vs. when you gave. Long term/short term capital gains and wash trade timing rules apply.

It gets pretty complicated when you have a trade where you are trading from a batch composed of multiple buys at difference price points in the past. Especially when you have to sort out the cap gains on the exchange fees you paid in crypto.

There is another YC company which helps you manage taxes for cryptocurrency trading. www.cointracker.io


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

Yes. You can define different tick intervals for different blocks. For example in one block you can select RSI with tick interval being 30 min and in another one you can select RSI with tick interval 2H. Ping on help button at the bottom if you get stuck. :)

Oh one more question - do you use walk forward testing? Monte Carlo simulations?

Not at the moment. But it is on the roadmap. One of the most demanding feature. :)

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

Thanks for the support. Looking forward to hear your feedback.

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/

We are going to add only reputable exchanges.

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?

There's a lot of algo trading projects on github - https://github.com/topics/trading-strategies is a good place to start

We have some sample strategies on the platform. We are working on building education material as well. You can also checkout out blog. https://medium.com/mudrex/tagged/learning

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

In our close beta 65% of the traders made profit. Usually people stop loss making strategies within a week or two.

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

You can signup and run backtest and strategies with virtual money for free. No demo amount is required.

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

We are working on adding more exchanges. It takes 2-3 weeks to test everything to make sure everything runs smoothly once live.

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

May I ask what platforms do you use for sports trading?

Betfair mostly, but also Matchbook, Smarkets and Betdaq.

Not on the roadmap as of now.

Great team! Excited to see where this goes.

Thanks. Looking forward to have you on the platform.

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.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact