Study Shows Forum Comments Can Predict Bitcoin’s Price with Over 80% Accuracy

[Nathan_Weymouth]

A research paper entitled “When Bitcoin encounters information in an online forum: Using text mining to analyse user opinions and predict value fluctuation” was published last year. Those interested can read the research paper here: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0177630#sec001

Do you agree with their finding that forum user comments can predict bitcoin’s price fluctuation with more than 80% accuracy?

[1714037878]

1714037878

[1714037878]

1714037878

[1714037878]

1714037878

[1714037878]

1714037878

[htsz_da12]

In fact, all price forecasts are guesses, all based on probability, the price band is not artificial, and has a lot to do with the market reaction and heat.

[dothebeats]

With the number of arbitrary predictions created at this very forum on a daily basis, I don't think the data gathered by the study holds weight and is not reliable. Sure, the analyzer would cherry-pick the words the researchers think is valuable in determining the actual price and place it in an algorithm, but at the end of the day it's just another study of guessworks translated to a pseudo-valuable research. Most of the time, predictions that come true are mere coincidences and not really because people knew that it was going to happen. The market always has the tendency to go against popular opinion, therefore making predictions inconsistent and an unreliable source of opinion for decision making and opening a position in the market.

[trecore4]

With the number of arbitrary predictions created at this very forum on a daily basis, I don't think the data gathered by the study holds weight and is not reliable. Sure, the analyzer would cherry-pick the words the researchers think is valuable in determining the actual price and place it in an algorithm, but at the end of the day it's just another study of guessworks translated to a pseudo-valuable research. Most of the time, predictions that come true are mere coincidences and not really because people knew that it was going to happen. The market always has the tendency to go against popular opinion, therefore making predictions inconsistent and an unreliable source of opinion for decision making and opening a position in the market.

Make sense. How come random predictions made by mixture of newbies and legends could average to a deciding scale of bitcoin prices ? I dont believe this paper at all and wont even believe it as long as it doesn't have people proof of work. I mean forum is good hub to know about many things, but predictions or analysis made here is just an uneducated guessing all the time. Most of the people here just type whatever they think about the bitcoin prices and many time they do spread FUD kind of stuff too. How can an online forum be so sure about the deciding factor of bitcoin prices.

[YuginKadoya]

We can not really rely on every speculation and prediction in the forum and some say that they are experts, It is just a coincidence that they have an 80% chance they are getting the correct speculation and that is the historical chart base on the previous movement of bitcoin, But that kind of theory sometimes can failed, Its just a matter of believing it because this is bitcoin and bitcoin does it all the time, And there are so many people believing because they have faith in the platform, or maybe they just don't have any idea about it and just going with the flow on the community which is sometimes dangerous because this is what manipulators are using.

[BrewMaster]

based on what i have seen in 2+ years of being with bitcoin and mostly on this forum and reddit i have to say this percentage is upside down. in other words 80% or even more of what you read on the forums are wrong. and when you are just analyzing them for statistics like this article you have no way of distinguishing trolls from FUDsters and real speculators and the forum is filled with them.

as a very recent example you can check the comments during June, if the 80% accuracy were real then price should have dropped below $6k and remained there probably around $3k that a lot of people were advertising for!

[wxa7115]

With the number of arbitrary predictions created at this very forum on a daily basis, I don't think the data gathered by the study holds weight and is not reliable. Sure, the analyzer would cherry-pick the words the researchers think is valuable in determining the actual price and place it in an algorithm, but at the end of the day it's just another study of guessworks translated to a pseudo-valuable research. Most of the time, predictions that come true are mere coincidences and not really because people knew that it was going to happen. The market always has the tendency to go against popular opinion, therefore making predictions inconsistent and an unreliable source of opinion for decision making and opening a position in the market.

I agree with this, if we understand the nature of the market then we will realize that this is a zero sum game, there is now way for everyone to win, in fact most have to lose so exchanges and speculators can get profits, so by definition the majority of the guesses and predictions have to be wrong, in fact in this market and in all markets it pays to be a contrarian, when everyone was buying at December that was the moment to sell and then when everyone was very negative about the future of bitcoin that was the moment to buy, now everyone is really positive about the decision of the SEC and this is telling me that selling may be a good option.

[andohyeb]

Forum members are the same traders that speculate on the price of bitcoins and I think their predictions are part of the reason that make the prices of bitcoins do as it does most times.

[figmentofmyass]

With the number of arbitrary predictions created at this very forum on a daily basis, I don't think the data gathered by the study holds weight and is not reliable. Sure, the analyzer would cherry-pick the words the researchers think is valuable in determining the actual price and place it in an algorithm, but at the end of the day it's just another study of guessworks translated to a pseudo-valuable research.

exactly. "cherry-pick" is the perfect description for this. i certainly wouldn't trust text-mining algorithms. i'm very skeptical that this kind of study can properly isolate the variables---they are probably reading too much into their deep learning model and not enough into a modicum of unstudied variables. like this:

Moreover, online user postings influenced Bitcoin transactions.

yeah, i don't think so!

[pooya87]

you know what i think, i think this is either one of these two things:
1. it is fake.
which means it is an attempt to make people think more of the comments they read on the forums so that they can be more successful with their FUD campaigns since they have been losing their effects.

2. it is misled.
which means they are not seeing cause and effect. they are analyzing effect and the follow up of that effect. for example when price is coming down, AFTER it started the descent there is a lot of topics/comments talking about doomsday and how bitcoin is falling hard and they start predicting the lower price. (same with the rises). they don't predict the drop then it happens, the drop happens then they start talking about it.

[davis196]

I guess that nobody on this forum predicted yesterday's price drop from around 8000 to around 7550 USD.
So many people here were riding the bull wave and thought "OMG,the price is going to skyrocket".
Last week's price rise was a"dead cat bounce" and the bear market will continue untill the winter.
The idea of predicting prices using the "wisdom of the crowd" is ridiculous.
The crowd doesn't have wisdom.

[Kemarit]

With the number of arbitrary predictions created at this very forum on a daily basis, I don't think the data gathered by the study holds weight and is not reliable. Sure, the analyzer would cherry-pick the words the researchers think is valuable in determining the actual price and place it in an algorithm, but at the end of the day it's just another study of guessworks translated to a pseudo-valuable research. Most of the time, predictions that come true are mere coincidences and not really because people knew that it was going to happen. The market always has the tendency to go against popular opinion, therefore making predictions inconsistent and an unreliable source of opinion for decision making and opening a position in the market.

Make sense. How come random predictions made by mixture of newbies and legends could average to a deciding scale of bitcoin prices ? I dont believe this paper at all and wont even believe it as long as it doesn't have people proof of work. I mean forum is good hub to know about many things, but predictions or analysis made here is just an uneducated guessing all the time. Most of the people here just type whatever they think about the bitcoin prices and many time they do spread FUD kind of stuff too. How can an online forum be so sure about the deciding factor of bitcoin prices.

Right. Looks like there's a lot of prediction algorithm being used, however, common sense will tell us that somewhere a lot the line, maybe it forgot to include one important factor, FUD. How do you rate a newbie account spreading shills around the forum or even a higher rank as well? And we all know that the market is based on pure speculation, which makes its harder to predict. I also don't believed in the results but I have to admit that the author has put a lot of time putting and crunching this numbers. But at the end of the day, behavior of the market and the price itself is unpredictable.

[coinplus]

A research paper entitled “When Bitcoin encounters information in an online forum: Using text mining to analyse user opinions and predict value fluctuation” was published last year. Those interested can read the research paper here: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0177630#sec001

Do you agree with their finding that forum user comments can predict bitcoin’s price fluctuation with more than 80% accuracy?

I really really really doubt that. I have been here for over 5 years now and I have seen people get the wrong price predictions like a million times so far. People that sees $5.8k just a month ago said it would go down to $2k as well by the end of the world cup, same type of people are now calling bitcoin to be $50k by the end of 2018 (which is just 4 months away) so I really don't believe forums can know the price with 80%, I doubt it knows the correct answer in like 10% of the time.

People online that are anonymous just gives out trigger amounts and make up huge numbers just to create attention. When bitcoin is $7500 and you create a topic saying it will be $7600 no one will care, it can go up 100$ and you can be right but that is just too small to care. However if you write it will be $50k than people will react to it and you get the attention you seek.

[kryptqnick]

In fact, all price forecasts are guesses, all based on probability, the price band is not artificial, and has a lot to do with the market reaction and heat.

That's true, but it might be that if many forum members keep predicting the same thing, rich newbies start believing in that and buy some bitcoins, thus making the predictions close to reality. I do agree though that there's too much data to analyze and, even though a lot was included in the research, we can't be sure it didn't omit other opinions on the matter. the research does look solid anyway and it is good that people try to conduct them.
It is not purely obvious what the 80% accuracy means. Is it that some kind of an average price prediction is 80% accurate, or is it that almost 80% of predictions are very close to reality?

[wxa7115]

you know what i think, i think this is either one of these two things:
1. it is fake.
which means it is an attempt to make people think more of the comments they read on the forums so that they can be more successful with their FUD campaigns since they have been losing their effects.

2. it is misled.
which means they are not seeing cause and effect. they are analyzing effect and the follow up of that effect. for example when price is coming down, AFTER it started the descent there is a lot of topics/comments talking about doomsday and how bitcoin is falling hard and they start predicting the lower price. (same with the rises). they don't predict the drop then it happens, the drop happens then they start talking about it.

It seems to me the second option is the most likely, members in the forum are trend followers, when the price of bitcoin was at 19k there were people prediction a price of 30k or even higher in the next months, then the price went down and people changed their predictions and some even predicted that we will get values of 2k or 3k, members switch sides very rapidly one day everything is perfect and bitcoin will change the world and the next day bitcoin is dying and the market will crash, so I really think that is what they are picking up with this study, so like always this is another study that puts the cart before the horse and confuses cause and effect.

[Tankdestroyer]

The idea of predicting prices using the "wisdom of the crowd" is ridiculous.
The crowd doesn't have wisdom.

The crowd does not have wisdom, but predicting prices is easier when looking at their opinions when compared to not having a general opinion to base your prediction on. At the very least, you can know what is the current sentiment of the public in the market and you can somehow predict what moves they will take sooner or later. Even if it is inaccurate, it is better than speculating without another basis.

[Pursuer]

The idea of predicting prices using the "wisdom of the crowd" is ridiculous.
The crowd doesn't have wisdom.

The crowd does not have wisdom, but predicting prices is easier when looking at their opinions when compared to not having a general opinion to base your prediction on. At the very least, you can know what is the current sentiment of the public in the market and you can somehow predict what moves they will take sooner or later. Even if it is inaccurate, it is better than speculating without another basis.

true only if you had a way of knowing who is lying and who is telling the truth!
if I say my prediction is bitcoin price will fall down to $3000 you have no way of knowing if I actually have that prediction or I want price to go from $7000 to $6900 so that I can buy bitcoin $100 below the market value.
you see it is not about inaccuracy, it is about people spreading nonsense in hopes that others follow their nonsense so that they can have the result they want out of the market.

looking back at the thousands of comments I have seen in the past 7-8 months that price has come down from ATH, a lot of them have been predicting ridiculously lower prices such as $3k I mentioned above by now. so if this "study" were true then price should have gone down there already!!!

[Vektrum]

A research paper entitled “When Bitcoin encounters information in an online forum: Using text mining to analyse user opinions and predict value fluctuation” was published last year. Those interested can read the research paper here: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0177630#sec001

Do you agree with their finding that forum user comments can predict bitcoin’s price fluctuation with more than 80% accuracy?

Very unexpected results of studies predicting the price of bitcoin, expressed in this forum. After all, these predictions are different, sometimes diametrically opposite in content, but in the end, 80 percent accuracy is obtained. However, everything can be. After all, predictions were made not just by the first people on the street, but by people who have some experience in cryptography.

[buwaytress]

PloS is a pretty good publication, for a free one, especially, and it does reinforce a lot of our suspicions: that the mainstream forum users here, despite what they would like to insist, are just as prone to sentiment influence as anyone else is. We see people say "don't be swayed" or "hodl on", but they panic just like everyone else when they see the FUD pop up.

That said, it's certainly logical that someone who comes out very rarely to say things will have more clout than say, a McAfee who's spurting our price predictions every 10 Tweets.

[qazgroup]

Yes i agree, crypto is a relatively new market and is highly manipulated, so we as a community have to be responsible because it does matter as the study suggests, so do not talk negative and always stay positive this will help and support the market growth as well.