Yeah it did mess that up - that was a pretty big change. A 7% change in the 24 hour prediction is pretty rare and I think that's the first change that far outside the average margin of error that I've seen. That will happen sometimes though, and trading based on these predictions may cause people to lose money sometimes. However, I believe it is correct more often than incorrect and that in the long run, if someone were to trade based on these charts they would end up earning money. That's why I'm testing the trade simulator project I describe in the Coming Soon page - so far it's earned a small amount of money but it hasn't had a lot of opportunities yet since prices have been pretty stable.

It's probably a result of the fact that the 20 day prediction has such a high average error. I wouldn't count on the 20 day prediction being particularly accurate. In all probability there will be some real-life events that occur in that time period which unpredictably change bitcoin prices. Those predictions are just a lot less reliable.

This is interesting, thanks for tracking it's degree of accuracy like this!

I wouldn't say this is the neural network's best work but it's not too bad I guess. It missed that bump in the beginning which probably threw it off a bit, but it seems to have done a pretty good job at other points in the graph. We'll see how it continues in the future!

Thanks! I hope the predictions serve you well

$607.39

at dawn

OgNasty/m3g4tr0n - yeah a lot of people have been asking about that, I think I'll try to make graphs showing this some time soon, or maybe extend the current ones to the left like you said.

segeln - Thanks for posting it to the German forum! Here's a bit of elaboration:

The part of a neural network that determines what results it will give (predictions, in this case) is "weights". A neural network consists of a bunch of nodes with edges between them, and the edges have certain weights. Initially these weights have random values. You teach a neural network by having it try to do calculations on data where you know the answer, and then adjust the weights based on how far off it was. So in this case, I have the neural network look at data at a few tens of thousands of points throughout the history of bitstamp and try to make a prediction. Initially the predictions are WAY off because the weights are just random. However, it sees how far off it was and adjusts the weights accordingly. Each time it trains on the historic data it is slightly more accurate. It takes about 25-30 ish training sessions for these neural networks to be able to predict with the average errors that I have listed on the site.

Let me know if you have any more questions!

rebuilder - Yeah I see it's predicting a pretty significant dip... I'm pretty nervous that it's going to be completely wrong lol but we'll see. I would not guess that it will dip down into the 530s or 540s but I suppose it's possible. I think what will happen is that it will dip but not by as much as it is predicting.

razfaz - Thanks! And that's pretty awesome if you were able to consistently earn 2-3% per month!

I use java for this project. I believe there are frameworks and libraries and stuff for neural networks, but I didn't use them. A few semesters ago I had to create a neural network for a class assignment, so I took the code from that and adjusted it accordingly to work for this problem (the old one was used for recognizing hand-written digits).

rebuilder - Okay I think we're not communicating clearly here, let me try to explain how it works as clearly as possible:

The neural network trains on historic data. The training process involves trying to make predictions at certain points in history and then changing parts of the neural network based on how far off the predictions were from the actual prices. The average error is how far off the predictions were once the network was done training. So basically the average error represents how far off each predicted price was from the actual price for predicting prices at every hour (or 6 hours, or 12 hours) in the past 3 years. I hope this is more clear, but please let me know if it isn't.

Cassius - I think if it just tried to predict up or down that it would be very accurate. I would guess that it would be right at least like 90% of the time, but I'm just speculating.

Cassius - That actually is how it works... the long-term previous data is used to 'teach' the neural network. I just haven't been having it output the predictions it makes while it's learning. I would be curious to see that data too though, so I think I will have it do that. Maybe I'll put up a graph showing this compared to the real historic prices.

Cassius - I'm not sure what happens because I don't actually see the predictions, but maybe I'll alter it so that I can. That would be interesting to see. I have a feeling it would predict very mild spikes and crashes at certain times but nothing to the same magnitude as the actual fluctuations.

segeln - I did it because I thought it would be interesting to see and other people had expressed interest in seeing it as well. I will definitely work on having it update more often too, don't worry xD

Hey guys,

So I've posted this here before, but I'm working on a software project that uses neural networks to predict bitcoin prices. Predictions can be seen here:
http://www.btcpredictions.com

A few people were asking me about predictions beyond 5 days, which was the longest forecast I had until today. I was able to get the software to make "reasonable" 20 day predictions. I say "reasonable" but really I would say they are probably very far from reliable. The average error is generally between 8.1% and 8.3% (meaning that if it predicts $500, the price will usually be between $460 and $540). So the 20 day prediction probably is not useful for trading, but it is interesting because it reflects the general patterns that bitcoin prices are taking over a slightly longer period of time.

And remember, the 20 day prediction is new and still very much in a "beta" stage. I will try to improve on it, but I can't promise it will ever be significantly more accurate.

GigaCoin - This method of prediction is not perfect (no method is or ever will be) but it definitely did call a drop in price, which did occur. It may have gotten some of the specific figures wrong, but the trend was correct (which as you said yourself, is what's most important). See the image below - the price dropped one hour earlier than the neural network predicted, and it reached its lowest point almost exactly when it predicted. I'll admit it was completely wrong about the rise in price that would occur after the dip, but I don't think you can just call this chance - this was accurately predicted.



Cassius - Thanks! Yeah I would like to add some longer-term predictions. I'll probably work on that over the next week or so... might even be able to get it up sooner.

dhomstad - haha yeah it is a bit weird. Creating those graphs was very tricky actually, making it so that it automatically adjusts the intervals and prices to fit to the data... I could work on making it have only 5 or 0 at the end of each interval and I might at some point. However, it's pretty low on my priority list... it is a lot of work for a small improvement.

segeln - Yeah you're right that the data is changing quickly. I might be able to make it update more often, but definitely not every one minute. The smallest interval I could probably do would be every like 10 minutes. Right now it takes the software something like 60-90 ish seconds to run, so it wouldn't be able to run every minute. It could theoretically do every 5 minutes, but I want to leave extra space because I may need it to make more calculations if I add longer predictions to the site.

Oh and I don't really have scientific background unless you count computer science lol... I'm studying computer science at college and I'll be graduating this May.

oda.krell - Thanks lol... and as far as your skepticism, it is not unwarranted. I haven't yet tested this a lot in practice. I do think you would be hard pressed to find a more reliable means of predicting prices though. So far it has been reasonably accurate, but I guess the true test will be when the price really starts becoming more volatile.

I believe that a neural network is capable of relatively accurately predicting prices when they become volatile - the question is whether or not this neural network can, or if I need to change the way it is structured/trained.

**Another interesting thing to note about the way this works though, is that over time this model will become more and more accurate. Bitcoins are still very new and there is not a lot of data to use for discerning patterns. As time goes on, more data will be available and so the neural network will be able to make a more informed prediction.

So yeah the price hasn't gone back up yet. Maybe it will, maybe it won't - the software seems to have changed its mind and is now predicting that the price won't go back up immediately. So the prediction was not entirely accurate, but I would say it was at least semi-successful. It was pretty accurate about the initial fall in price.

well I hope you are right lol

Well I just woke up and looked at the charts - it looks like the price definitely started dipping down around 3am (1 hour earlier than I predicted) so I'll call that half of the prediction a success. It has yet to rise back up again but we'll see. Like I said, it may get the magnitude wrong or it may shift slightly in one direction or the other. The price seems to be behaving reasonably close to what the model predicted though.

lol dude don't you get tired of trolling?

P.S. plz tell me more about ur hot gf

Haha I will call the prediction successful if it happens within like +/- 3 hours maybe, but no more than that. Certainly not 20 haha

Torque - This is a computer science project, I don't need to know a lot about economics to program a decent neural network. If you dislike my project then ignore it, there is no need to come here and insult me and my work for no reason. At least give constructive criticism if you don't like it - that I would be open to.

segeln - thanks man!

Torque - I don't know what a shark trader or whale is, I don't know what a FUD article is, I don't know what half the things you said even are. I spent hundreds, if not thousands of hours working on the software that creates this, and I never said that I have some amazing predicting powers. In fact, I have been continually warning people that this can sometimes be wrong - there is no guarantee here. You obviously did not read anything I wrote on my site. Please don't make ridiculous accusations towards me based on literally nothing -_-

Also I'm not working with anyone else on this - it's just me at the moment.

Actually neural networks are very commonly used for stock predictions and they are known to be successful. You can buy software specifically for this purpose.

Check it out: http://en.wikipedia.org/wiki/Stock_market_prediction#Technical_analysis

Also I never said they are magic... Like I said, my software is sometimes wrong. It does seem to do pretty well though.

ShroomsKit_Disgrade - Thanks!

MOB - The software doesn't have a name because I didn't name it, and the neural network is an artificial neural network, not my brain. http://en.wikipedia.org/wiki/Artificial_neural_network It's a computational model for approximating functions basically.

Cyberlight -
1. It's not always right, as I said in the OP and as it says in multiple places on the website. It probably performs better than a human could though.
2. It did predict for the price to go down, just not to the degree that it did.
3. The average error is the amount off from the actual price that each price usually is. That means that with an average error of 1.3%, if the prediction is $600 then the price will usually be between $592.2 and $608.2. I explain how the average error is calculated in the about page.
4. The idea that I created this for market manipulation is beyond ridiculous for far too many reasons to name here...

*Also if something in China caused the price to go down, as what you said implied, then there is no way the software could predict that. I explain this in the section about real-life events on the about page.