Incentives Seem Off


Maybe I am off, but the incentives for this contest seem to be misaligned. As in the contestants are not properly incentivized due to small rewards.

For the contestants that are actually good at making these predictions, why would you stick around here instead of trading in your own account? Brokers like Interactive Brokers provide a lot of decent data.

What am I missing?


I was coming here to make this exact post. I’ve worked in hedge funds for years and know what a decent financial engineer makes. Why on earth would you put in this kind of effort for a fund that probably manages in excess of 1B USD and pays out just a few thousand for a top model? It’s really really way off.


Neither of you have looked at the data yet? There is no way to make any sense of what the data actually represents.

This is a pure machine learning contest, no financial background required. I’m an astronomer, others here are biologists or geologists, and for us, this competition is perfect to practice some machine learning techniques on a really difficult problem. If this dataset was similar to any other data set provided by places like Quantopian, then a lot of people would not even attempt to take part because they don’t have the financial training or background.

We don’t submit models to Numerai either. We only submit our predictions. Most of the people here don’t do this as a day job, we do it for a chance to learn some machine learning techniques that we can use in our day jobs. This is a side project for most that take part.


Themicon, OK those are valid points. You have no interest in dealing with the trading side of this. The data is fun to play with and you actually get solid results, which is refreshing to some.

From my point of view, with some background involving trading and ability to deal with the operational side of trading, I have no reason to be here. Sounds like Kevbot is in the same boat.

I leave you gents and ladies to it =)


The main disconnect is probably in that you seem to be overestimating the amount of effort involved. Obviously I don’t speak for anyone buy myself, but I can spend about 30 minutes a week, plus a few hours computation time and have a decent chance at pulling in a few hundred extra dollars a week in the staking competition.

I very much doubt I could get the same result by actually trading myself as I’d most likely need more time working on models plus a high quality data source, so probably more investment in terms of both time and money.


Daenris, I see that point as well. I think this just comes down to a difference in goals and perspective.

Coming at this from my point of view, as someone that wants to continue investing with their own funds, my time is better spent developing trading strategies and executing on them.


Yes, absolutely. Numerai’s competition ultimately has nothing to do with investing or trading from the side of the participants, and so your initial question just seems to make no sense to me, because from my point of view the rewards are not at all small based on the amount of time I put into the competition.

So if what you want to do is invest on your own, of course it doesn’t make sense to participate in Numerai.


The incentives were good in the beginning. Or when NMR was worth way more than 11$.
There are still incentives there to make a little extra pocket money; it may not be balanced though. Given the time wasted for the originality test to pass and the number of tries I have to go through everytime to make a successful submission, the incentives are going out of focus every week.
My estimate is that if NMR goes down, and Numerai doesn’t make up for it with rewards in USD, we might see less competitors.