Providing a detailed description for payment and staking

One problem I am encountering is understanding how to get paid. I’ve downloaded the dataset, read through the docs, and looked through the sample models. As I understand it, the way to get paid is by producing the most relatively accurate stock market predictions.

However, the problem is I don’t understand the staking system. For example, Numerai gives you 0.1 to start and I can stake that, but the scoring page and staking & payouts page are unclear. Moreover, what happens if I lose that 0.1? How do I get more? Can anyone provide a more concrete description for the payout system and incentive structure?



I too amstruggling to understabd i have staked the free 0.01 but now i guess i have to wait and watch the performance of that model over the next 4 weeks… But im very confused regarding how to stake more and if its even worth staking on my model tbh lol

1 Like

Brian, welcome to the community. The most common advice I’ve seen from numerai ‘elders’ is patience. I am fairly new too, and started building up a stake after two months of submissions. From what I’ve read about how a model on a winning streak suddenly starts burning for months when market regimes change, two months was probably too early, but I couldn’t resist.

The way you get more NMR is to purchase on an exchange (binance, coinbase, bilaxy and uniswap seems to be mentioned the most).

The incentive structure is: you risk real money, your profits and losses are determined by how much you risked (and your prediction accuracy), and numerai relies more on predictions from people who put skin in the game for their hedge fund.


The way to acquire a significant amount of NMR to stake is the same way (more or less) you’d acquire Bitcoin (BTC) or Ethereum (ETH) – from a crypto exchange (like Coinbase, etc) NMR is a so-called ERC-20 token, which means that it is “made of Ethereum” in a sense (it exists on the ETH blockchain). This crypto side of things is complex and confusing not only just because it is, but also because how easy it will be to sign-up to an exchange will depend on where you live and its crypto and banking regulations. You don’t have to find an exchange that supports NMR directly – if it does not then you can just acquire ETH first at your exchange (they all support ETH) and then get your NMR from Uniswap (swapping ETH for NMR). But one way or another, in order to buy NMR (with real money, fiat like US dollars) or eventually sell NMR back to real money you’ll need some service/exchange to do that which will probably involve verifying your id, hooking up your bank account, etc. And there might be tax consequences too! Welcome to crypto!


Thanks all for the comments, I really appreciate your time. I don’t think the incentive structure is where it needs to be for the tournament to be worth the time investment.

It is definitely overwhelming at first as there is a machine learning/data science component, a staking component, and a crypto component. Crypto is very confusing to most everybody at first. And there is a double-risk component – you can lose some of your stake (under current structure would actually be pretty tough to lose your whole stake, but it is technically possible) but the price of the NMR token itself can go down so there is risk there. (And it has been jumping up the last week or so probably mostly because of all you new users which makes it kind of a bad time to buy some – maybe, if it just keeps going up it is a good time, you never really know.)

Which is why it is recommended to do one thing at a time. Unless you already active with crypto I would make that the last of your concerns because if the machine learning part of things is not going to hold your interest then no point in even going there. So I’d be thinking about making models and seeing if that is something you like doing and whether you can make any good ones.

As far as is it all worth it in terms of incentives? Well, it is definitely risky, no doubt about it (especially with that added currency risk). BUT…the trends have been good. The price has been up & down before (it has been up to $60 something in the last year for brief periods while spending most of its time in the $20-$32 range for many months, and only a year or so back it was about $5). And the better users are making quite good returns, like really really good just in NMR terms, on top of the fact that the fiat value of that NMR has doubled or tripled or quadrupled for most of them.


No. I understand machine learning and blockchain technology well. But, Numerai is pulling Tom Sawyer’s trick here by making developers pay to do the work. I looked at their data and it is relatively useless.

Thanks for your insight though, I appreciate your time. Best of luck.

If you couldn’t understand how to get more NMR (a token available on many exchanges plus Uniswap) it didn’t seem like you did understand it so well. And if their data was useless then their hedge fund would suck and none of our models would perform well. Except it doesn’t (although detailed info here is restricted I admit) and many models do well (the data on this is quite clear). And they’d really be getting ripped off as I understand the cost of that useless data is enormous.

Hey, it certainly isn’t for everybody, but the upside is large and it is real.


I am not sure what the hedge fund’s performance looks like because the information is not publicly available. But, the Numerai dataset won’t support the best returns, which typically require slightly more risk than hedge funds are willing to take on. Moreover, even assuming the hedge fund’s success, it can be attributed to the virtually free labor. So yes, you are getting ripped off. There are better publicly available datasets, or you could make your own with a web scraper.

Nothing personal, just a fact. You should get paid to work. Therefore, Numerai would be better if the tournament rules were more cogent and open ended. For example, producing the model with the best returns for a certain month will yield a fixed rate in return.

I was saying if their data is so terrible, they are getting ripped off in buying it. (The same data other hedge funds buy.) Anyway, you are getting into hard-to-take-seriously territory by making a bunch of pronouncements about how things are after being here 5 minutes. So if you aren’t interested, you aren’t interested, and leave it at that.


No problem. That makes sense. Thanks again for the input, I really do appreciate it.



If you are still around @btrain69 , consider this. Numerai is using a metamodel approach. If you were going it alone with your own data and models, chances are your volatility will be hard to tolerate. Maybe it’s not hard if you are young and can always earn your money back in your late 20s, early 30s. But later on you will start to become more risk averse because of pets/kids/house/spouses/old parents etc. Not to mention whole economies including the global economy work best when volatility is controlled.

To put it another way, the underlying assumption you seem to be working with is you will be smarter than everyone else. Maybe - it’s human to think you are smarter than everyone else - it’s also arrogant. Even if this week you look like you are the smartest person in the market, there’s a good chance sooner or later your model is going to fail and when it does your arrogance is going to make you pay. Hedge funds hate this because their investors will be on the phone as soon as the fund drops 10%. So, the metamodel uses a ‘wisdom of the crowd’ approach. Maybe the predictions aren’t the most accurate, but they are also more likely to not be very wrong, week to week.


Hi Scott,

Thanks for the note, you’re a great writer. Additionally, your research in NMR and computational biology is fascinating. I apologize if I came across as arrogant because it was not my intention. I sincerely appreciate the valuable feedback I’ve received on this post and I am happy to continue the dialogue.

So, I’ll respond directly to your points regarding hedge fund risk, the metamodel, and going at it alone. First, if your hedge fund is based on a blockchain architecture, then you’re not looking at a 10.0% drop. In fact, total failure manifests in many myriads. Consider, the SEC decides to prosecute the fund, along with its investors for unregulated securities trading, like they did with Ripple Labs in December 2020. Or perhaps, simply civil litigation arises and IBM sues the fund for patent infringement and every dollar they haven’t made yet. Either way, a hedge fund based on a blockchain for controlling market volatility is not a method for reducing risk.

Next, on the metamodel, the ‘wisdom of the crowd’ doesn’t disguise the heard mentality, nor the associated pitfalls in capital management. Most people probably buy into the idea that more data and more models yield more money. But a contrarian truth reveals, you need to think different to succeed as an investor. Which leads me to the last point, it’s better to go at it alone. The main reason is because it allows you to cultivate your own data, rather than using Numerai’s garbage data. I call it garbage not to be arrogant, but because the data is literally worthless in an encrypted form. To most successfully build a financial model, you need to know the variables and you need to mine the data yourself. In doing so, you can use your unique knowledge to gain an unfair advantage against the market.

In conclusion, I’m not paying to compete in a tournament where I can’t use my knowledge to my advantage because that would take away my edge. Short term, I’m making better returns in the market on my own. Long term, I’d rather invest in something substantive, like starting a biotech company because owning equity in a company that succeeds is the best way to create both personal and social wealth. Again, I really appreciate you taking the time to respond and I am happy to talk more.



@btrain69 Maybe checkout Numerai Signals?

1 Like

That’s helpful - thanks!

Hi @btrain69

I think you raise some good points. I’m thankful for the thoughtful discussion! I apologize for the long essay I have written below. Its a public holiday today but I still have child care for my child and nothing pressing at my real job. Writing like this helps form my views better so I sometimes throw up on the page as a way to do that.

Firstly, I’m sorry if I sounded like I was suggesting you are arrogant. That would be a failing of my writing style. Suggesting arrogance wasn’t meant to be personal but rather just accepting that we all are arrogant in our opinions. This is a human failing which impacts us on an individual level first and then others second. But this is my point regarding the metamodel. I’ll return to that in a moment.

Your point about crypto is well taken by me. I’m personally very bearish on the whole crypto market even though I’ve made some money in it. I mean who could not with a simple buy and wait strategy? But as a system, its future is uncertain, and as you correctly point out it’s legal issues are real, which means I predict high volatility in that space. I am quite risk adverse, which I guess is starting to become obvious. I’d prefer that crypto was not a part of staking with Numerai. The Hedge Fund is making real currency backed by real nations that have a vested interest in continuing to exist, while we make cryptocurrency backed by the blockchain and it is less certain that that system will have the same level of desire to continue to exist. This point is controversial and no one knows for sure. But we (the tournament participants) do take on an extra level of risk than our masters - it’s not fair - but in the end so be it. It’s an extra level of risk I can tolerate. But I take this position because I’m here more for the datascience challenge and less so to make money. I doubt I will make substantial amounts of money anyway and I’m not looking to. So we may have different purposes here that lead to different opinions.

Regarding the metamodel I think we disagree a lot. I’ll try to justify my position again. I can accept that the ‘wisdom of the crowd’ (WOTC) philosophy can fall into heard mentality. But 1) an individual investor is not immune from this either 2) the metamodel can solve this problem so long as it is composed of guesses that are as independently formed as possible. Only a metamodel of some kind can solve this type of problem. An individual can not. Even if they think they are avoiding the heard mentality, it’s impossible for an individual to do so. That last sentence is opinion, not fact - I don’t believe anyone has facts either way. So, how can WOTC suffer heard mentality as well? Simply by having participants pay too much attention to what each other is doing. Does that happen here at Numerai? Certainly it does. WOTC does not solve the problem with algebraic precision.

But I think Numerai is using WOTC in a way that helps solve the problems associated with WOTC. And again we disagree a lot here. You use the term ‘garbage’ to describe their data - its a strong pejorative the word ‘garbage’ but I understand better now what you mean. I’ll use the word ‘obscured’ instead. I maintain that obscured data is a strong way to prevent heard mentality. Lets say ‘Feature X’ is a stock’s price correlation with Dogecoin. And let’s just say that stock is GE (no particular reason, just making things concrete). We notice that this feature has had a consistent value over a few eras. Now Elon Musk makes some pronouncement about Dogecoin. If too many people worship Musk (sadly too many do and I think a lot of them might be here in Numerai) then many will ignore their objective models and insist that Dogecoin movements as a result of Musk statements will impact GE price. Decisions like this have so many levels of bullshit baked into them it needs its own paragraph to discuss.

Firstly, Musk knows next to nothing about what makes a stock price go up or down (IMHO). Even if he knows something good or bad about Dogecoin or GE, and this would have to be insider information (whenever you see insider information, think volatility and run away - you are about to get burnt or imprisoned), his statements are exactly what will make the correlation fail and the feature to become incorrect. If you believe something external about the feature you will hypothesize about the feature. An investor will make mistakes because they know what the stock and feature is. But he will also have to fail to understand the difference between correlation and causation. A model uses the features to guess the target by finding correlations between features and targets and if our model is smart we find correlations between complex nonlinear combinations of features and complex nonlinear combinations of targets. Nonetheless it’s all correlations. Correlations are fine - but if we know stuff about the features and how they reflect the real world we naturally start building a hypothesis about causation. But to establish causation you need to run controlled experiments and you simply can not do this robustly on the macroeconomic scale. Causation (not correlation) will lead investors wrong because we can not establish causation. But OK, maybe you think you wont be (mis)lead by interpreting your knowledge of the features and targets beyond their raw values and you wont try to establish causation. Then why have any information about the features and targets other than their raw values? This information about features/targets is exactly what will lead to heard mentality because, for example, too many people trust Elon Musk or try to establish a causation between Dodgecoin and GE without any basis in reality. Obscured data is the remedy.

In short, we get in our own way.

But as @wigglemuse pointed out, maybe Numerai Signals is more your thing. I want to explore that too at some point. But again, going it alone using your data in isolation (your own silo) is IMHO risky. Sure, use your own data, but I don’t recommend making investment decisions in isolation. Even if you are the smartest person in the room you will blow up one day and maybe for an entire week and maybe until you have no money left. It would be nice if you were buffered by a dissenting opinion. You see, even when you are the smartest person in the room, a dissenting opinion might slow you down, but one day that dissenting opinion will stop you from blowing up. And it’s important that that dissenting opinion be based on methodology as different as possible from your own. I think obscured feature/targets in a metamodel achieve this goal best.

Again, I’m enjoying the discussion. I think I will try and put some numbers behind what I have just said and write some simulations of how I think siloed opinions blow up. I don’t think that is original work but it will make things more concrete for me and maybe convince you and others about the wisdom of the metamodel. Maybe not :rofl:

Robbo (the Fossil)


Thanks Robbo. Your point about Elon Musk is valid. With that said, page 23 in the latest Tesla 10K is worth reading because he points to the SEC’s selective enforcement scheme as the reason for investing $1.5 billion in Bitcoin.

Your point about currencies backed by nations is well taken. It is important that the U.S. Government has a significant interest in blockchain’s success because they own at least $10 billion in Bitcoin, probably more. As a system, for a while I was most attracted to Warren Buffet’s description that cryptocurrency is rat poison squared. But now that Bitcoin is worth more than both Berkshire Hathaway and Tesla, it is too big to fail.

The cryptocurrency market is now worth $1.29 trillion. Looking forward, Coinbase, which recently announced plans to IPO, is the most valuable asset in the digital economy. So, consider Coinbase will be worth more than Apple by 2024. Which leads to my last point, Numerai is most valuable insofar as they are the cryptocurrency with the smallest market capitalization listed on Coinbase at $194 million. So, it is also one of the most volatile cryptocurrencies because it is most subject to institutional manipulation.


All I have to say is “herd”, not “heard”.


LMAO - You got me @wigglemuse

Oh dear me - 5 times I did that! Wow - just wow! At least I am consistently wrong.