NMR Price Appreciation

In chat and here in the forum, folks have discussed the value of NMR; its day-to-day price movement mostly, but sometimes there have been attempts to value it using different theoretical frameworks. In this post, instead of rehashing those calculations or sketching out my own theoretical estimates, I want to just summarize actual recent historical price movements, say over the past year or more.

I will but briefly say just here at the outset, that all of the reasonable theoretical discussions (I guess you could say, those that I have been able to follow) point to a somewhat bleak picture for NMR. The reason for the worry is that NMR price seems to be supported mostly by purchases of NMR by data scientists in the main tournament. In the meantime, payouts are going down due to the continued decrease of the payout factor. The main motivational effect of the payout factor, in my experience, is to make it so that I don’t have to worry so much about my models. This means that we can expect model quality to degrade. Again, IMHO, the amount of NMR held by staking data scientists (under exponentially decreasing payout factors) will always be so small that it cannot be thought that an increase in their number will somehow support NMR price stability or even price appreciation, but instead price variations will be fueled mostly by weak speculation on the open market. In other words, there is really very little incentive to buy NMR on the open market. If this is the case then we would expect NMR to weakly trail market movements dominated by more powerful tokens. I think that is bad news for staking data scientists.

OK, so having painted a bleak theoretical picture, let’s look at the data and see what it says about that theory.

The following plots compare the token-to-BTC price ratio (normalized to one at the starting time of each plot) of the top 15 coins on Coinbase by market share (that have at least a year of data, except MATIC which comes in just under a year) – to the NMR-to-BTC price ratio. The time scales are 1. a little over 4 years. 2. 1 year, 3. YTD, 4. 6 months, 5. 3 months. 6. 1 month.

1: ~4 years

2: 1 year

3: YTD

4: 6 months

5: 3 months

6: 1 month

To explain the relative price variations here briefly in case you dont want to have to stare at the plots: At the ~4y timescale, NMR is at the bottom of the pack. At 1 year it is still at the bottom of the pack. At YTD it seems like it starts to want to move (2nd from the bottom). At 6 months it is making some headway at 4th from the bottom. At 3 months it seems like it wants to take off, even surpassing DOGECOIN in 6th from the bottom. But at one month it has fallen firmly to the back of the pack of 16 tokens. The market is firmly saying that NMR lags; it’s nothing special and that nothing that we are doing (we being Numerai, the CoE or the community), is working to make it special in any way.

So question is can something be done to make NMR special? I think the answer is yes, but the theory and the data suggest that it will not be by continuing the present course of actions; I think really radical changes are needed.


There is one major difference between NMR and most other coins.
NMR generates good old dollars through the hedge fund.
NMR generates value in a tradition real world business ( if a hedge fund is a real business :slight_smile: )
As long as the fund is successfull, there will be incentives to keep NMR alive at least with a stable exchange rate or better.

Other coins don’t generate value in other real world busineses. I guess most will disappear with time, because they don’t generate value. They are kept alive by the crypto mania.
And it takes time for others to realize a real world use case for NMR, simply because most tokens don’t have any.

1 Like

I am glad you brought that up; that is the key and it would be great if it were true. You see, Numerai has done everything it can to distance the value of NMR from Hedge Fund performance. Go ahead and ask anybody on the team. The issue is that if NMR is related to the Hedge Fund performance, then NMR could be considered a security – that has implications. Well, that is the argument that I have heard. So, every time I have asked I have been told that there is no relationship.

Also, if you look at the data, you can see that there is nothing to indicate that what you are saying is true, at least up to now; NMR price is trailing other coins in the crypto market place. Something remarkably different would be happening if what you are saying is true.

1 Like

In legal term, yes you are right. NMR is not a security and there is no legal relationship between NMR and fund performance. That is clear.
And yes, you are right, NMR price is highly correlated with bitcoin. I guess because most people think about it as “one more coin”, with no intrinsic value.

However let’s assume that the Numerai fund has a great performance based on the predictions we are submitting every week. It produces real dollars for the fund. Depending on the size of the fund that could be a sigfinicant amount.
If NMR price keeps falling due to some external factors (like a collaps in crypto markets), then data scientist would leave Numerai as well, hence killing the fund performance and the profit that the fund generates. In this situation Numerai is incentivised to spend some real dollars to keep NMR price at a reasonable level. Otherwise the fund collapses as well. For that to happen, of course the fund needs a high enough AUM. I guess we are there yet.

Again let’s assume, that the fund has $1B AUM and the fund gets 2% annual management fee.
That gives Numerai $20M every year to fund itself and keep NMR afloat.
That is enough the buy up any sold NMR every day. NMR liquidity is less than that on most days.

It’s still hard to calculate a fair value for NMR.
My point here is, that NMR has some intrinstic value, that can be expressed in dollars! Even if the market doesn’t recognize it yet.
Most coins can’t claim that!

1 Like

NMR is an asset the can generate USD through the fund.
If something generates USD then it must have value that can be expressed in USD.

If the fund performance is 0, then NRM value expressed in USD should be ~0.
As long as the fund generates USD, NMR value is should remain greater than 0 USD.

You have control over the fund with the NMR you stake.
Your reward (divident) is a function of your contribution to the fund and a function of overall fund performance on the long run.

It’s hard to deny that NMR is much like a security.
Not exactly a security in legal terms, but something similar.

By the way, market is always wrong.
That’s why I could retire at the age of 37 :smiley: and that’s why we can achive positive correlation with our models.

1 Like

I might be wrong this time…

1 Like

The most important thing the CoE can do with regards to token appreciation is to determine the end state status of the security once the treasury runs dry. If NMR is not considered a security, then 0 is removed from all present value calculations. This dwarfs a million BS content proposals in overall effect on NMR.

The argument being used by the team against Numerai purchasing NMR is that NMR will be considered a security if Numerai uses fund profits to purchase NMR. Numerai was able to purchase BTC in the early days, and large companies now are purchasing BTC funded via debt offerings, and BTC is still not a security. So there is a very prescient legal question to be worked out there for NMR and crypto in general, and CoE should find this out on our behalf immediately.

The other specific assumption that needs to be validated is, once the treasury is depleted, it is assumed Numerai will purchase NMR on the open market to pay us. Again, I do not see how treasury level determines Numerai’s ability to purchase NMR on the open market.

The fund profits >> coin mechanism (at any point in time), is critical to NMR price appreciation, and a simple legal review would provide investors with a better means of determining a present value.

With regards to data scientists leaving the tournament due to NMR price, I don’t think it really matters to the fund at this point. The meta-model correlations are absurdly high for most participants and the meta-model would ultimately just consist of whatever the team is able to think up and incorporate (which would look a lot like the meta-model today) if the tournament ended tomorrow.


I think it’s much more useful to look at the long term prospects of NMR, and Numerai in general, than to worry about the short term fluctuations.

The basic model for NMR, at least as I understand it, means that when the MM is successful, then payouts in NMR increase the circulating stock of NMR, hence decreasing its value relative to other commodities.

When the MM is not successful, the circulating stock of NMR is reduced through burning, hence increasing its value relative to other commodities.

Personally I think that aspect of the model is brilliant.

There is the question of how long Numerai can keep making payouts from its treasury. My own ballparks estimate, based on the size of the treasury and the amounts paid out on good rounds, is on the order of 10 to 20 years. I’m comfortable with that :slight_smile:

So what happens when the treasury run dry? Well, at the current payout rate, I’ll be long gone, so I don’t really care. But if Numerai is still going, they’d have to buy it on the market, which would tend to raise the price of NMR. If that cost exceeds the profitability of the MM, then Numerai goes under, otherwise they can continue if they like.

As the payouts decrease, will that discourage data scientists from participating? It might discourage the bad ones, but if you have a solid model, and you automate your process, then after the initial ramp up you should be able to make a steady return with very little effort. It’s sort of a Darwinian project in that respect.

1 Like

But this is exactly what is denied by the Numerai team. This is the crux of the matter. And unfortunately, I have to believe them more than I believe your argument. But see the discussion by @of_s.

1 Like

Not necessarily. The payout factor roughly reduces the effect of burns as much as the gains. The effect is not exactly symmetrical because during burns some folks pull their NMR out. Either way the reduction is enough that staked NMR will persist much longer and model efficacy has more time to be worked out with less frequent effort. I think that in general this means that folks will not bail so quickly and weaker models will persist longer. So we can expect the meta model to decrease in quality as the payout factor continues to decrease.

1 Like

My suggestion from day one was to expand the use case for NMR, it is both early days and some use cases are getting crowded but Numerai once said they wanted to control all the money in the world or something like that, ok what about a DEX ? that seems to be in line with the mission, a stable coin and other defi products like lending/CDPs could also expand the use case. I think we are heading into a world where traditional securities and other finance products will be mirrored if not entirely traded/offered on chain, so Numerai can benefit from having early products. Another use case could be financial data since they already spend a lot of money getting it, why not source and provide data in exchange of NMR.

Unfortunately I think it needs to come from them since these are not trivial products and would need their team to expand/legal considerations etc, etc, but what we as users can contribute might be limited.

1 Like

This may have the effect to give NMR more liquidity. But order for a token to appreciate it should become more scarce while liquidity actually makes a token less scarce. So the effect of liquidity on NMR would probably make it even more like other cryptocurrencies and cause it to lag even farther from mean token-BTC price ratio. Some other action is needed IMHO.

Use case expansion should really come from the VCs who have a huge vested interest and the connections & economies of scale to facilitate this well beyond what us participants could.


I agree that the solution probably needs to come mostly from Numer.ai itself. There may be a use for the CoE in all of this but see the discussion by @of_s.

Has the MM decreased in performance as the payouts have decreased? I have no idea, and it would be interesting to keep an eye on.
But in any case, if the MM performance were to decrease because of large staking by bad DSs, then that makes an even greater opportunity for good ones.

And we should take compounding into account. Suppose a good DS (GDS) and an equally bad DS (BDS) start at the same level. GDS gains (say) 25%, and BDS loses 25%, and they let it ride. GDS will then have 66% more influence on the MM than BDS. It’s a wonderfully self correcting system.

I’d say that MKR for instance burns tokens when you pay CDP fees, so NMR could be burnt that way, a DEX basically has fees that it divides between pool holders and itself, those fees could be paid in NMR, so yeah scarcity should be a concern.

First of all I have two problems with your graphs:

  • You should compare them in log scale, because multiplication is what matters.
  • It is not fair to compare NMR which is less then 200th coin by marketcap with coins in the top 50, especially after the big correction. Because lowercaps crash more, and recover later. Largecaps are always first movers.

My take on NMR price. I 10x my capital on Cardano last year, and now NMR is my largest position. To find gems like ADA you have to be lucky or work pretty hard. I invested like 8 hours every day to research coins to realize that Cardano is the best bet, and I still think it was partially luck. To gain big, you have to invest in coins before the information you know is known to general public. For me NMR is a safe bet. It grows faster then BTC, almost nobody knows it, and I can gain big stake rewards.

Why I think NMR has and will have value.
Markets are not rational. If they were, I bet they would be already solved.
What is value and how do you define it? IMO value is about trust. As long
as we can trust Numerai team, NMR has value. It’s like Bitcoin with utility.


Idk, I guess for investing there are better choices, like easy-to-use Defi products. I would not invest into NMR. I guess the price of NMR increases if the tournament is paying nice rewards and more people are staking on their models, so there is demand. However, I’m not sure how many people have those skills. Also, it has to compete with staking-rewards of ~90% APR from pancakeswap, raydium or cake defi and potential growths of those coins.

EDIT: It might sound that I don’t trust numerai, that is not what I think. I think the rewards are definitely there and the whole project is amazing. However, I’m not convicted that the risk/reward-ratio is good for people who don’t stake on models. If there is a way for everyone to stake on models, the price will for sure explode.

Point 1: On every plot the starting point is renormalized to one. If you did the same thing on a log plot you would not see much difference because \log(1 + x) \rightarrow x for small x. But what is more important here is the relative position at each timescale, not the exact numbers.

Point 2: This is exactly my point. If NMR had some intrinsic value beyond just another crytographic token it would not do what you are arguing that a small token does – just lags the big guys. The graphs clearly show that it lags; it does not have any intrinsic value beyond what the market values a not well known small cap token.

Was there an increase in NMR supply recently?

I see the market cap increased but price remained the same.