Buy and Hold or Buy, Hold, and Sell?

>>2984085
Hold. it's flatlined. Pump is imminent any time this month.

>>2984085
sell first, then keep selling.
thank me later.

>>2984085
the idea is good.

but the execution is flawed.


stay away from this is my advice.

File: 1499630356427.jpg (100KB, 433x419px) Image search: [Google]

100KB, 433x419px

Still holding $1200 worth of bags at .0475

>>2984263
Same

>>2984097
This seems to be the general attitude of everybody holding NMR.

That being said I'll still hold and see what happens, I will get my money back one way or another.

I refuse to sell them for cheap, I'd rather rode them to 0 then give in to these sell horde mentality.

>>2985003
but the devs said its not for long term investing. you will never get back what you lost

>>2985041
That's a blatant lie. And that's what biz told you.

Number.ai cut NMR rewards by 90% to encourage people to hold. They even said so "number.ai's new economy"

It might never get it back but I refuse to sell it then. Guess that NMR will be lost forever.

>>2985100
whats to stop them from cutting the rewards more?
or one day all together ?

>>2985100
but I read the whitepaper
It exists purely for data scientists to create algorithms using the numerair market data

>>2985227
Let's think about this for a second. Why would you want numeraire if it is worthless?

Why would you stake numeraire if it is worthless?

You would not care about earning numeraire in this case. Since it has no value in this case. Why would you bother trying to even obtain it?

To prove this point, 90% of the weekly payout is in numeraire, because they want to give it value.

They could just pay 100% of the rewards in
BTC and let you participate without numeraire. The more accurate you are the more cash you get.

So why don't they then? No matter how you look at it. Numeraire needs to have value to
be wanted and for the system to work. There is no way around this.

Numeraire is only useful if it has value.

>>2985118
Cutting the rewards more or all together would make NMR even more valuable since no new NMR is issued. And they do plan to have a hardcap of around 21,000,000 for NMR.

The sole reason you could bring up would be: "What if they cut you out and just stopped the project using the AI the scientists made"

But this is also a flawed reasoning, the A.I. might perform well now but might not when the circumstances change.

In this case people would be distrustful of joining numer.ai again which would come to bite them back in the ass.

The A.I. will need constant adjustment to function as it should.

So numeraire needs to have value for this entire to work, no matter how you look at it.

File: 1-1z1eesVCQDnK6Xi4vXagQA.jpg (207KB, 1595x758px) Image search: [Google]

207KB, 1595x758px

>>2985374
they don't need numberai tokens to operate the hedge fund portion of their business , ai cannot predict consistently better than 50-50 the market unless they are also using value investing techniques.
but we have no idea what or how they are investing, since they obscure the data.

so many potential cluster fucks.

>>2985392
>ai cannot predict consistently better than 50-50 the market unless they are also using value investing techniques.

Proof of this?

File: randomwalk.jpg (31KB, 227x346px) Image search: [Google]

31KB, 227x346px

>>2985405

has nothing to do with the AI, its just the nature of the market.

https://medium.com/towards-data-science/how-i-predicted-the-stock-market-at-numerai-ml-tournament-6f74e1c8809e

http://www.thesimpledollar.com/review-a-random-walk-down-wall-street/


its not correlation- just random time sudo accuracy


The accuracy score gotten here is 0.514361849391. Pretty low so I decided to raise the value of the C to 100.0 for better classification and high variance. This actually took almost forever(15 mins) for the classify to run but gave a score of 0.518133997785 which is just a bit higher right than the previous score. It’s still low but I’m glad to know it’s a bit above average. Fair enough for a newbie :)

>>2985405
tl;dr , its basically a thinly disguised lottery for data scientists .

their probably not gonna wanna keep losing money to this sort of lottery over time so people stop submitting data.

i dunno. seems like they fucked up the execution of this coin. seems close to something that could have been cool tho . oh well .

File: 1500016490329.png (10KB, 436x439px) Image search: [Google]

10KB, 436x439px

Science tastes funny

Flawed coin or not, I still think there's a lot of people watching it it after its huge 10x in a day run and if it starts making a run again, people will jump on it to not miss out. Doesn't have to work to get pumped.

File: 1501251183578.png (303KB, 800x497px) Image search: [Google]

303KB, 800x497px

>>2985503
sure.

but i do know they're are much wiser investment strategies than to hope for things like that.

and u can always use your nmr to try for the competition if u are any good with python or w/e so there is that.

i personally believe having computer try to predict random short term market moves with no other news or company information is insane but w/e
but like i said
holding bags could be a fun incentive to fuck around with python

>>2985534
I don't hold any now but I do watch it every day. If I see a spike I'll hop on and ride it, that's about all it is for me.

Numerai has big potential for anyone bothered to look

>>2985629
care to explain why you believe that to be true , and perhaps even refute some of my points against it if you want?

>>2985687
Numeraire.ai was created in the first place to try and create an artificial intelligence that can predict the market.

Saying it is not possible at the moment is why the project was founded in the first place.

The reason they are using an AI is because an AI can see patterns that other have failed to see or are to complex for a human to grasp. A working model could have 74 dimensions of inputs for all we know.

That's the whole premise of the AI in the first place. If you do not believe this is to be possible. Then yes, you have no reason to invest.

If they start to get a working A.I. it would be in their best interest to buy up some number.ai to keep the price high so more people can pile in a better can be achieved.

File: 1501903066836s.jpg (13KB, 250x249px) Image search: [Google]

13KB, 250x249px

>>2985801
>If you do not believe this is to be possible

it actually could be, using the proper investment techniques.

but those investment techniques don't require artificial intelligence so.
lol . wtf .

if they were churning thru a database of all possible stocks trading at bvp higher than price
and buying them, and selling as soon as price was lower than bvp , doing this high frequency , that would be a very desirable model, but it honsetly seems like, from how they interpret the data set, they want the data scientist to try to predict a coinflip .

its very gross actually , a mis appropriation of intelligence.

but hey, m aybe they're marketing will be powerful enough to keep market theory ignorant data scientist to gamble on their platform

>>2985374
what the fuck? It IS worthless. This is the crypto craze

>>2986051
Yes anon you keep thinking all crypto's are worthless. In the mean time I will keep buying more.

>>2986010
It is clearly not, there is not solid way to predict the market. Because if there was one, by consequence there would some entity that would be slowly usurping all the money. Which is exactly numer.ai's goal by the way. To own all the money in the world.

The reason AI is being used is because the problem is of such great complexity that no definite or even consistent method has been found working as to date. AI is a black box, it thinks in ways a normal human would not think. It would give more credit to some things while classifieng others as completely irrelevant. It is the ultimate way to solve problems, you give it inputs and tell it to maximize certain outputs.

Its the same way a human brain works but without any external influences or ethics being involved. The AI only cares about performing well. No emotions, ethics, morals, gut-feelings matter. It only cares about performing well.

I strongly believe in such a model performing well given enough time as recent advancements in AI allowed such robots to beat humans in complex games like Chess, GO, ... Even now Google is working on an AI to beat Koreans in Starcraft. The potential here is limitless.

File: GottGoDown.jpg (157KB, 800x500px) Image search: [Google]

157KB, 800x500px

There is but only one direction.

>>2986656
when the fuck did I say all cryptos are worthless?

and if the Numerai AI is so good, why is coin always hitting new lows? Does this not strike you as a typical scummy hedge fund?

File: make-it-rain-o.gif (4MB, 320x261px) Image search: [Google]

4MB, 320x261px

>>2986656

>The potential here is limitless.

if you say so, i'm pretty sure, based on the data sets i have seen that the fund distributes for the tournaments, all the hedge fund is doing is applying the closet coin flip predictions to short term buys. in hopes to flip a quick buck.

at least they destroy the tokens if the coinflip python data turns out to be "wrong"

they aren't feeding the data to some super computer ai machine , lol.
if they were they could just do the calculations themselves in 'R'.


i need you to understand how fucked this whole scheme is.

so feel free to continue to engage me until you are satisfied this is a complete birdbrained waste of money.


https://github.com/dbjohnson/numerai


its a tournament to see who can guess a coin flip the best based on past data they then verify.

that is all.

>>2986743
explain to someone who only knows basic calculus

File: makeitrain.gif (1MB, 350x197px) Image search: [Google]

1MB, 350x197px

>>2986796


what part of this is a dev scam casino coin, not meant to make any real money for anyone but the hedge fund owner ,and the completely

arbitrarily lucky "data scientist" money?


u get a set of a bunch of bullshit numbers.

you try to write a program you think finds a correlation slightly above .5
*50%
50/50 coin flip.
on the data set.


you submit your program , and money to participate in the tournament (ie : lottery)

then your code is run on a completely different dataset.

the code that just so happens to correlate the highest above .5 (50/50)

on the new dataset, wins the tournament(ie: lottery)


the only thing stupider would be , if the hedge fund actually did take this "algorithms", and try to use them in the actual stock market.
but since the founder is a math major , and obviously financial enthusiast., i have a feeling he knows better than that.


its a lottery.


programs cannot find whether a stock will go up or down short term, ESPECIALLY without any underlying financial fundamentals information.

>>2986796
https://numer.ai/rules

its like writing a computer program to guess "heads or tails"
10 times.

then taking that computer program, and
paying money to see if it can guess heads or tails
out of 10
more than the other computer programs.

does that make sense?

its bullshit.
a computer cant guess heads or tails,
ofc truly random number generation is impossible by a lay computer , so there will be a slight bias
after enough attempts.

its bullshit. its not fucking SKYNET

its a very complicated casino game. nothing more.

if you still want to speculate, do it based on that knowledge, not their bullshit futuristic marketing nonsense .

>>2984085
>Bought at $78
>Dumped at $160

You guys are cucks stop buying this shit

>>2986869
no

Isn't the number of NMM decreasing over time? It should increase in value then.

>>2986667
Ye it loves to go down

File: 1499454801856.jpg (56KB, 598x766px) Image search: [Google]

56KB, 598x766px

>>2986906
aren't the number of polio cases decreasing over time?

polio should increase in value then.

>>2986675
I have no clue why it keeps going down in such a consistent matter. I suspect heavy price manipulation is at play here.

The names behind numer.ai are also well known, I doubt they would want to blemish their reputation for a few millions.

>>2986960
the tokens value isn't tied to in any way, the value of the "hedge fund"

and even if it was,
the fund would at best break even (without considering fees) if they are trying to uniformly apply those data sets to the real market for profit.

>muh data scientists
so much vitriol for this coin from biztards....sure sign it's gonna moon hard at some point....nice, just bought 100

>>2986743
There would be no point in doing this. Of course they are feeding it to an AI. Why would you bother letting random people control your hedgefund? There is no point if it really is as "random" as you describe it to be.

The whole project would be pointless in this case. People might as well bet on stocks then, it is called a hedgefund for a reason. It is ment to make large bucks. Not to let some random scientists bet whether it will go up or down. Who would want to invest in that? Nobody.

The reason they destroy the tokens is because they do not want unsuccessfully data scientists to supply them with wrong models. This would be a waste of time.

This AI is not what you think it is. There are no real "calculations" being performed or anything you could make sense of as a human.

They are training some massive neural network to set this thing up. They are not trying to distill some polynomial function.

>>2986848
Number.ai takes NO taxes like a lottery does, so that would be pointless.

You can also participate for free in tournaments, only staking requires NMR. Why would they give out FREE MONEY unless they used the data models.

Lotteries makes sense if you buy a ticket. There is no ticket to buy here and if you loose the NMR gets destroyed. Number.ai gains 0 of this yet they give free BTC if your model works. I wonder why...

>>2987003
>>2986991
>>2986980

if you say so,
i'm just telling you i think its kinda bullshit and scammy.

>>2986848
>>2986826
unironically made me think

>>2987052
>>2987030
I have no clue whether numer.ai will succeed or not. But they are not giving away money for free and it's not gambling or a lottery either for data scientists. That's just incorrect.

>>2987188
I believe I read on the white paper that it is a lottery pool for the data people

>>2987188
you just dont understand the math behind it.
thats fine.

maybe it will be successful based on that pervasive ignorance.

idgaf , im over it.

you dont understand the math behind my explanation. lets leave it at that.

>>2987222
I don't need to understand the math since there is nothing complicated about it. You are citing some basic chance calculation which has absolutely nothing to do with the goal numer.ai is trying to accomplish.

The fact that you or comparing the stock market to coin flips is what makes you seem retarded here. Those things are completely different. The data is obfuscated but are essentially certain indicators/averages from the stock market. Not some randomly generated data.

Again, number.ai is trying to find a pattern for the stock market or for the data you are given.

"Proof of Intelligence" is key here. Not proof of fucking randomness/luck.

Jeezes fucking Christ what is so hard to understand here?

You act like some smug retard but understand nothing of the project. What a neural network AI is or why these data scientists or giving models.

It is NOT A LOTTERY. Since the NUMBERS ARE NOT RANDOM.

Do you get it? Or will you come with some new explanation on why it is purely chance based?

You can treat is as much as a lottery as you want and enjoy loosing your NMR because you are creating random bullshit that does not actually work consistently. Numer.ai wants a model that works across multiple models and will give you numeraire in return.

I am saying this stuff is not a lottery and numer.ai is not a charity nor a damn lottery company.

Jeezes Christ you are dense.

>>2987550
wait a second... pattern.. stock market

thats impossible

>>2986946
Wow you're an actual retard. Judging by the tripfag name I'm guessing you'll be going back to school soon...Jesus I hate summerfags.

>>2987550
Did you even bother to read the white paper you retarded fuck. It literally said in in the first paragraph but here you are spouting bullshit.

"A common approach to verify accuracy in machine learning is to break the dataset into train and test sets.
A trained model can be tested for accuracy on the test set, which it has never seen. However, to maintain
statistical validity, this test set should only be used once. When a data scientist accesses the test set multiple
times and uses that score as feedback for model selection, there’s a risk of training a model that overfits the
test set. This hurts the model’s ability to perform well on new data."

Here you fucking go, now shut the fuck since you clearly do not know what you are talking about.

>>2987582
That's what they are trying to find by using some smart AI. That's the point of the project