Honest Mirocana ICO review

Honest Mirocana ICO review

Nikita Kolmogorov
This is an English translation of the infamous Russian article I wrote yesterday that got me banned at the official Russian Mirocana Telegram group. More than a dozen of people were interested in receiving a translation so here it goes, enjoy!

Well hello there, dear reader! I had some spare funds the other day and decided to invest them into a promising ICO at the pre-sale stage to increase the chances of getting the largest ROI possible. A friend of mine shared the link to the Mirocana ICO with me and I decided to investigate whether it was worth to invest and if I could trust the founders. In this article I will try to review pros and cons of choosing Mirocana tokens as your primary long-term investment.

ICO is somewhat like IPO with the grain of the blockchain technology. IPO stands for Initial Public Offering when large enough companies (Facebook, Microsoft, Apple, etc.) start selling the shares. ICO is almost the same but instead of buying a share of the company control and receiving dividends, investors either pre-order goods or services or literally receive a speculation tool that may (or may not) rise in price over time.

Foreword

Why shouldn't you trust me? Because I created the product that directly competes with Mirocana, my own algo-trading platform; hence, the presence of a conflict of interests and preliminary scepticism. So take everything I write here with the grain of salt. Please, check all the facts I present here before jumping to any conclusions. Whether to invest money to Mirocana or not is totally up to you, I just tried to save you some time by investigating their ICO myself.

Why should you trust me? Because I've been working in the algo-trading field for almost 6 years. Because my code is displayed as a code example at Wex.nz (ex BTC-e.com) — take a look at the Objective-C part. Because I've been doing technical consulting of various companies for almost 4 years now. Because I've been a part of more than 50 projects — usually, participating from the beginning till MVP launch. Because I've launched over 10 personal projects (the most recent ones: voicybot.com — over 1 000 000 users, arbeitbot.com, @banofbot, Controlio — currently closed, Bitcoin-draw — currently closed, trading platform — my main source of the revenue at the moment). I've probably seen enough in cryptocurrencies, startups, investments, development, marketing and scaling.

Till this moment you have probably already decided if you want to continue reading or drop this article. If you decided to keep on reading, then make yourself a cup of tea, cover your legs with a warm blanket, sit beside the chimney, relax and keep reading. Let the review begin!

Mirocana

Let's try to gather as much information about the product as we can: what is it exactly and why would users need it. Mirocana presents itself as an automated tool for securities, currencies and cryptocurrencies trading. Soon you'll see why this complex tool might be the perfect one for every investor out there. The main 4 products that Mirocana promises to deliver are:

  1. Core product — automated bot trader itself, which can trade using your money until you run out of the tokens bought during the ICO. It consists of the parts: securities, currencies and cryptocurrencies trading, and you get access to each of them depending on how much tokens you've bought.
  2. "Alpha" — allows users to create and test trading strategies.
  3. "Sigma" — allows users to create and test machine learning trading algorithms.
  4. "Target" — neural network that should predict whether an ICO is going to be successful or not.

Altogether, we would be able to make the machine to increase our deposit, test our trading strategies (the ones we wrote), test our machine learning algorithms (the ones we wrote) and predict the success of any ICO.

But why won't these products launch or satisfy the customers?

In my opinion, the Mirocana team is trying to do too much and simply isn't familiar with the MVP concept. Instead of focusing on the neural network that, according to the feedback of developers and one of the users at their official Russia Telegram group, make over 80% of profit a year, they spread the efforts on the gigantic separate features that are large enough to be considered separate products. Let me explain.

MVP stands for the minimal viable product — a product that has just enough features to start bringing value to the users and charge them for it.
The classic image representing what MVP is

"Alpha" and "Sigma" will require to either use existing programming languages to describe the trading and machine learning algorithms or will require to build new programming languages from scratch. And new programming languages aren't easy to write from scratch, it is a well known fact by anybody who has ever taken programming languages courses at any university. I repeat: building a new programming language from scratch is difficult and very time consuming.

The issue is that describing machine learning and general trading algorithms are two completely different tasks that will require two different programming languages with two different styles and even paradigms. It means that the Mirocana team will likely have to build not one, but two programming languages from scratch.

But let's assume that the Mirocana team will take two existing programming languages to save some time. They will still have to establish a high-load web-service which will check trading and machine learning algorithms. Frankly, it's not easy as well. If you can make client-side languages to evaluate trading algorithms, when it comes to machine learning algorithm evaluation it gets very tricky.

While training my own neural network to trade cryptocurrencies I used my own data which I gathered since 2011: bitcoin prices, open orders, market volumes, completed trades, etc. The size of the data scares me every time I have to get the hard drive with it out of the cold storage. And, in order to teach the neural network, I had to rent a supercomputer for a couple of days. In order to complete the process the neural network went over this gigantic array of data billions of times.

How will Mirocana give users access to such a testing opportunity? I don't know. How can we debug the algorithms then? Will we be able to slightly tweak the algorithms and run them over and over again consuming days of super computational power per platform user? Obviously, it can't be done client side and Mirocana will have to probably spend major part of their budget on building the servers capable of running the neural networks over terabytes of data billions of times per couple of day per user.

Moreover, Mirocana probably doesn't even have the access to the data array large enough to efficiently train the neural networks. I tried to buy such data on most of the opened and closed market around a year ago to have a dataset different from mine to better train my neural network, and can proudly say that I'm among a few people on Earth having this kind of information — in particular, opened orders every 5-60 seconds starting with 2011.

But even if they don't need such amount of information to train the neural network, think about how many people will be required to develop such monsters as "Alpha" or "Sigma". Let's get minimum numbers: 2 data scientists per project, 2 front-end developers per project, 1 system administrator per project, 2 server developers per project. 14 people in total minimum. Obviously, I just made up these numbers and you don't have to trust me but it's a reasonable number for the project worth $40 000 000 before active development begins. But those 14 people are just developers — what about PR, marketing, development management, QA? It easily gets up to 50 people. And managing a company of 50 people is very different from a team of 11 people in Mirocana as they state on their website.

And if we take into account the fact that they also need to develop the core product and "Target", the company size might easily go up to 100 people as well. But "Target" is the whole another topic. There were only 247 ICO's so far. What do you think: is it enough to train a neural network? Yay! I think the same: it's not enough. And taking into account that every ICO is exceptional, it will be pretty hard to figure out a general rule or algorithm for this neural network.

But the core product — it works for 1-2 years, as the developers say! It trades their money and the deposit of the secret investors pool, details of which have never been disclosed. The only way to check if their neural networks actually brings profit to users is the bot on Telegram that should show us opened and closed positions. This is the ultimate proof that the algorithm works and can make money, right? But this bot hasn't been working properly recently and developers have no time to fix their core product — they need to spend time raising money (official public statement of the Mirocana CEO).

Official answer by the Mirocana CEO

Then I found an article written by the Mirocana CEO on a Russian startup website vc.ru about Mirocana bot. Probably, their first publication and it included the link to the real-time Mirocana trading results! Unfortunately, the results end on April 3rd 2017 (maybe bot stopped working then?). You can see for yourselves that bot hasn't made much profit with the initial deposits at $10 000. They've earned 10% at max on average and a major part of the accounts lost money.

This screenshot is attached just in case if the page will be removed or altered — it would be a pretty bad move, by the way, and believe me, I'll notice

All in all, Mirocana promised investors everything that they could only wish existed on the financial algo-trading market. But everything is alright! There are CEO's that can effectively manage such companies on Earth, there are people with enough management and development expertise to build the profitable product for $40 000 000 even if they have nothing at the moment. In case if they are experienced, of course.

But let's take a closer look at the Mirocana team

Everything will be alright if we have genius and exceptional managers, developers and traders! Let's see who we got.

CEO, the founder, the core of the whole project — the person who should lead the project and the team toward the profits by his exceptional management and development skills! George Petrov is his name. 21 years old guy who has thoroughly cleaned his social network footprint and left us with his Facebook and LinkedIn.

His Facebook is pretty boring, so let's jump straight to his LinkedIn (the best blogging platform of 2017) — for sure he should have tons of relevant experience in management and development! Let me walk you through his career:

  • Python+Django Developer @ Twingo, April 2015 — June 2015. Summer internship? Part-time? 2-3 months of working in a company as a Python developer? Not bad, but I bet he wouldn't learn much during this short period.
  • Head of product development, co-founder @ SafetyFirst, May 2014 — September 2015. Company website is down, but it feels like this was an average apps development shop. But why going to work to Twingo at the same time? I'm not sure. How many people did George manage? I bet, less than 5.
  • Backend Python Developer @ ZAO "Otdel", September 2015 — December 2015. Company website is down as well. And again, 3-4 months of Python developer. Taking into account that this company has already dissolved, George likely didn't manage more than 5 people as there were only 7 developers in total, according to his LinkedIn.
  • Python Backend Developer and Data Scientist @ AMT-Games, 2016 — 2017. I suppose, here is their company website. George writes that he worked with the mobile fames server and was building the machine learning algorithm for the match-making system. The issue is that it is an entry-level issue that any senior developer can solve within 1-2 months. What did George spend a year there on — I don't know. Did he manage anybody? I don't know, but I guess he didn't.

Where is the management experience of teams with 15+ people and extensive machine learning experience, Lebowski? George claims that he started programming when he was 13. I started when I was 12 and by 20 years old I've already had my own mobile development studio with over 15 employees. Thank God, it's over now and the studio was closed in 2013. And what? Can I build a profitable trading algorithm out of nothing? For sure I can, but I spent 6 years building it. Who's next? We are interested mostly in the developers.

  • Evgeniy Ulyanov — Head of Operations. Here's his LinkedIn — usual PR specialist who talks to the people. Not a developer.
  • Daria Patanina — Head of Marketing. Here's her LinkedIn that is completely empty except for Mirocana project. This is the main marketer. For the project, trying to raise $40 000 000. Obviously, not a developer.
  • Vasily Boychuk — Head of Data Science. Here's his LinkedIn — worked less than a year at Motorola in 2013, in 2015-2017 worked at a Russian university. What did he do at Motorola? What did he do at the university? What was he up to during 2013-2015? Frankly, Vasily has no relevant GitHub trace. I guess, he has never participates in projects worth $40 000 000. He is a developer.
  • Alexey Dementyev — Head of Product Development. Here is his absolutely empty LinkedIn. Can't say much about him. Facebook is empty as well. I guess, he is more like a manager, rather than a developer.
  • Update from January 2nd, 2018: *name removed* left the company in October, 2017 — Mirocana still hasn't removed them from the website.
  • Dmitry Tolstyakov — Head of Strategies Division. Here is his Facebook, where he announces his investment webinars. Here's his website. No visible connections to Mirocana. I asked him on Facebook about his connection to Mirocana, as soon as he responds — I'll update this part. Might be the person responsible for the trading algorithms. Not a developer.
  • Vladislav Pochukalin — Head of Design. No contacts provided. Can't google him. Previous experience can't be found. I guess, he is the freelancer who made the design for Mirocana. Great job, but he is not a developer.

And here it is — dream team for the service worth $40 000 000! Two programmers (one is the founder) and 6 managers with questionable experience. On the official website they claim to have 11 people in their team. Which means, that aside of this wonderful eight we have at most 3 more developers. In total, we have 5 developers (one of them cannot spend time programming — he has to run business) who need to build an auto-trading platform for $40 000 000 based on the Telegram bot that has shown poor results. And why didn't they include the rest 3 team members in the White Paper? Where is so called transparency, Lebowski? And why those 5 developers cannot fix the main indicator of the core product working correctly?

The wonderful eight

So, would you like to give $40 000 000 to this team for bare-boned promises? And that's not all folks, keep on reading!

We are moving to the advisors

Let me remind you that advisors are the people who actively advise and consult a startup. You can't call people whom you wrote: "let us give you $3000 and put your name and face on the advisory board" — advisors.

What advisor does the automated neural network trading platform have? Any guesses? Data scientists? Machine learning specialists? High-load specialists? Financial gurus? Economists? Well-known traders and investors? People, who actively advise and believe in Mirocana? Yikes!

4 advisors left after the scandal

Moreover! A pretty interesting situation happened when I found a post on the Facebook where the author and commentators not only uncover fake advisors at Mirocana, but also present proofs that Mirocana tried to "buy" advisors. The story goes as follows: initially Mirocana had 10 advisors and 10 research specialists, all of them top 10 in their fields. But the author wrote directly to 5 of the research specialists and they haven't even heard of Mirocana. Later we learnt that Mirocaca has been adding advisors to their advisory board without their consent.


Well, I wrote to all 4 remaining advisors:

  • George — received a prompt response, consults Mirocana on the ICO process. Doesn't consult them neither on technical side nor on trading.
  • Denis — also received a prompt response. He says that he consults Mirocana only on the internal finances. Not trading or investments, but only internal accounting.
  • Update from October 23rd, 2017: *name removed* is no longer advising Mirocana as he confirmed in PM on LinkedIn. He has sent them the notice. I repeat from his words: "No longer involved, due to time constraints. No funds of any kind were ever exchanged between me or George. I can’t speak to anyone else. I also never accepted anything. Nor will there be any funds. Between George, Mirocana and I".
  • Andrei — I also reached him on LinkedIn and also asked George to ask Andrei to answer me. According to his profile he advises Mirocana on the legal problems. But not trading or investments.

To be completely objective, Andrei and *name removed* haven't responded yet. As soon as they respond, I'll update this part of the article. Also Mirocana claims to have 2 more secret advisors that haven't yet gave their consent. But for some reason I find it hard to believe that they are actively adivising Mirocana on tech or investments.

But the bare fact that they've used fake advisors might mean that there are no relevant advisors in this project at all. And it is usually one of the first signs of a scammy ICO.

Initial list of the scientists that has already been deleted from the official documents. At least 5 of them declined working with Mirocana

I've read their boring White Paper and around half of it is dedicated to the ICO process, a bit about the product itself and nothing at all about the existing Proof-of-Concept. No details about the current implementation at all (except for the obvious ones).

But there is a very (very!) cringey moment in the end of the White Paper. They claim, that Mirocana Tokens do not qualify as securities in Canada. However, Canadian OSC has officially stated that ICO tokens that can be sold later to other people are qualified as securities. Moreover, they claim that the token price is going to go up in the White Paper itself! And keep in mind that Mirocana is registered in Canada! Even though you can spend those tokens later at Mirocana web-service, it still can be traded for money which qualifies it as a security. And it doesn't matter how many times and in how many places will Mirocana write that their tokens are not securities, Canadian law is still above Mirocana.

Fragment from Mirocana White Paper

Mirocana is registered in Montreal, Canada. Which means that the first inquiry about Mirocana that will get to OCS can lead to company closing down and assets arrested. Nice ROI, isn't it?

Update from October 20, 2017:

Mirocana team presented the "Legal Opinion" written by the IPS law firm, can be downloaded here. It contains legal opinion that Mirocana tokens aren't securities in USA. Unfortunately, no words about Canada. Moreover, there is a typo in the very first paragraph ("Mirocana Tokets") that raises some questions. Even more: in the end of the legal opinion, inside disclaimer, they write that this paper cannot be used in any way but personal advice about whether Mirocanas tokens are securities or not. It also says that this paper is just an professional opinion that can differ from reality. It would mean that Mirocana cannot use this paper for commercial reasons.

We still wait a solid, well-written, without typos, legal confirmation from a Canadian law firm about Canadian law. Not the legal opinion paper, but actual real legal confirmation that the law firm that wrote it can stand for it. And, yes, IPS is the company within the conflict of interest, so we need a third-party legal audit.

Presumably legit legal paperwork

Conclusion

Well, what can I say, here's what we've got:

  • Telegram trading bot that probably stopped working in April but even while working has shown poor results
  • 21 year old CEO without enough management experience (let's assume that he has enough development experience)
  • 5 developers for gigantic platform consisting of 4 parts, each of which might be considered as a completely separate huge product; them not being able to fix the only proof that neural network works having 5 developers in house
  • Fake advisors history, no relevant advisors, proofs of the attempts to "buy" advisors
  • Team, most members of which don't have any experience working in a middle-sized company on a top-management position
  • Potential risk of legal issues

So what: to buy or not to buy? I think that if you are going to invest then do it only to speculate and sell the tokens at the peak. I guess that there will still be plenty of people who'll invest without reading this article so you can make profit on them. I wouldn't suggest to buy these tokens for long-term because as soon as the scam will unfold and the platform will start showing its inefficiency, investors will start selling tokens like crazy.

Please, note, that I have nothing against George (the founder of Mirocana). Even more: i'm glad that such businessmen appear in Russia! I fully support George. I hate when people start their careers with scam.

I hope that Mirocana team will crash every single point of mine: product, team, advisors and legal parts of the project. It will be great if they show us how product really works on realtime data (i.e. posting trades in real time on the Telegram channel) — simulations can always be faked. It will be great if they have team members with enough management, development and machine learning experience. It will be great if they have advisors that actively advise them on machine learning, blockchain, high-load, and can verify active participation in Mirocana. It will be great if they present true legal confirmation that their tokens will not be considered as securities. Until then, we'll wait.

Thank you for reading such a long review! If you feel like you'd love to support me — just follow my page on Telegram — @golden_borodutch. It's solely in Russian though, sorry :(

--------------------

Updates from October, 20th, 2017:

  1. They banned me at the official Russian Mirocana Telegram group becaused I asked logical questions and presented arguments agains Mirocana. What do their PR specialists do? They are digging their own grave with such methods. I guess, this happened because I declined talking to them in private messages but asked to talk to my in public so that everybody could see.
  2. *name removed* (one of the advisors who could finally be the one advising on tech) responded on LinkedIn. He advises Mirocana only on VC relations. He doesn't advise them on tech or investments. Mirocana officially doesn't have no relevant advisors on board.
  3. One of the subscribers sent us screenshots of how Mirocana Telegram bot traded since May 2017. It made +94.95 for the $10 000 deposit, or +0.95%! Yes, less than a percent of profit in around 4 months! Keep in mind, that real-life trading differs from theoretical simulated trading because of the errors occurring on markets and in bots. For example, my platform has shown -30% decline in revenue after I launched it to trade real money as compared to the same amounts, but simulated.

Update from October, 21st, 2017:

  1. They banned me at the official English Mirocana Telegram group. The only thing I did was posting this article — nothing else. I haven't even asked any questions. Looks like Mirocana tries to hide this article instead of responding to it.

Update from October, 22nd, 2017:

  1. Mirocana CEO has promised to release the official response before the end of October 23rd, 2017.

Update from October 23rd, 2017:

  1. *name removed* is no longer advising Mirocana as he confirmed in PM on LinkedIn. He has sent them the notice. I repeat from his words: "No longer involved, due to time constraints. No funds of any kind were ever exchanged between me or George. I can’t speak to anyone else. I also never accepted anything. Nor will there be any funds. Between George, Mirocana and I".

Update from January 2nd, 2018:

  1. *name removed* left the company in October, 2017 — Mirocana still hasn't removed them from the website. This person reached out to me with the request to remove the name.