Author

Topic: Wall Observer BTC/USD - Bitcoin price movement tracking & discussion - page 28151. (Read 26720865 times)

donator
Activity: 1722
Merit: 1036
This is a very low R^2. You are predicting 5.7% of the outcome (possibly because bitcoin's walk is primarily random).  So what this means is that all of your predictions will have huge error ranges.  With a large error range you might predict a 10% gain on average, but actually it will range from -30% to 40% and may have only a 60% chance of being greater than zero.

If you don't believe me now because of a low R^2, why don't you believe the exponential trendline which has R^2 = 0.94?  Wink

Yes, I know it's very low but in this kind of game where you can use leverage if you want etc., every edge is an opportunity. Unless it is purely random, which I don't believe. I intend to run the same analysis for different lengths of time, perhaps to find better results.

One of the fascinating things about Bitcoin is that a random month is positive with 60% probability and only 40% is negative.
legendary
Activity: 1596
Merit: 1030
Sine secretum non libertas
Keyser is no troll, as you well know...what happened to you Aminorex?

To be clear, I was describing the post, not the author.  (One may be known to others by one's deeds, but we are not identical even to the sum of our deeds; far less are we characterized properly by any one deed.)
sr. member
Activity: 364
Merit: 250
I wish someone would take a few months to gather all the predictions made in this forum and make a list of the best nd worst predictors

 'Someone' has already invented such a prediction-validating mechanism. In fact, this very thread was originally created for speculators who engage in such predictions.

 It occured in very rudimentary forms since ancient times, saw a further sophistication during Roman rule, disappeared during the dark ages (obviously), reappeared in the late middle ages mostly around Italian city-states, & finally, emerged in its modern financial trading market form in ~1602 in the Netherlands, as the Dutch East India Company's shares were first traded among investors.

 Later, the ~1880's featured the advent of something yet closer to current currency/crypto trading exchanges - world currencies began to be backed by gold, and the first large-scale retail forex trades started taking place.

 It wasn't until 1996 that the Internet gave these foreign currency exchanges the capacity to run the entirety of trading electronically. Various digital new currency projects such as e-Gold/etc were soon undertaken excitedly by a series of hopeful entrepreneurs.

 ..Sadly, in large part due to vulnerable centralization & the absence of math-based strong encryption systems as core features of these innovations, they ended up failing & dimming public opinion of any digital currency concepts as irreedeemably flawed Ponzi schemes. Nevertheless, amongst the ashes of these unsuccessful experiments, a different, vastly more sound, secure & viable project came to light : Bitcoin.

 It is thusly that in early 2011 the very first bitcoin exchange markets were founded, bringing us full-circle to the present day as we find, with the warmth of serene satisfaction, that exchange markets indeed offer the best mechanism through which to filter the least skilled from the most skilled predictors, by comparing traders' accounts' initial investments with their net equity several months or years later.

(Sorry, I just love history xD)

Indeed.  In fact, we have a blockchain that can provide audited records to show who has the best results.  If only we had a way to link the BTC addresses in the blockchain to posters here.  Ah, but we do.  A signature can provide that proof.
hero member
Activity: 910
Merit: 1003
I found this mildly interesting, an inside glimpse of MtGOX in 2012:

  March 29, 2014 - Reuters
  Exclusive: Mt. Gox faced questions on handling client cash long before crisis
  http://www.reuters.com/article/2014/03/30/us-bitcoin-mtgox-idUSBREA2T01T20140330

Sounds very similar to the Neo & Bee story (but of course the amounts involved were 50-100 times bigger, and the drama lasted years instead of months).
full member
Activity: 175
Merit: 100
I wish someone would take a few months to gather all the predictions made in this forum and make a list of the best nd worst predictors

 'Someone' has already invented such a prediction-validating mechanism. In fact, this very thread was originally created for speculators who engage in such predictions.

 It occured in very rudimentary forms since ancient times, saw a further sophistication during Roman rule, disappeared during the dark ages (obviously), reappeared in the late middle ages mostly around Italian city-states, & finally, emerged in its modern financial trading market form in ~1602 in the Netherlands, as the Dutch East India Company's shares were first traded among investors.

 Later, the ~1880's featured the advent of something yet closer to current currency/crypto trading exchanges - world currencies began to be backed by gold, and the first large-scale retail forex trades started taking place.

 It wasn't until 1996 that the Internet gave these foreign currency exchanges the capacity to run the entirety of trading electronically. Various digital new currency projects such as e-Gold/etc were soon undertaken excitedly by a series of hopeful entrepreneurs.

 ..Sadly, in large part due to vulnerable centralization & the absence of math-based strong encryption systems as core features of these innovations, they ended up failing & dimming public opinion of any digital currency concepts as irreedeemably flawed Ponzi schemes. Nevertheless, amongst the ashes of these unsuccessful experiments, a different, vastly more sound, secure & viable project came to light : Bitcoin.

 It is thusly that in early 2011 the very first bitcoin exchange markets were founded, bringing us full-circle to the present day as we find, with the warmth of serene satisfaction, that exchange markets indeed offer the best mechanism through which to filter the least skilled from the most skilled predictors, by comparing traders' accounts' initial investments with their net equity several months or years later.

(Sorry, I just love history xD)
legendary
Activity: 1870
Merit: 1023
Shameless cross-post to advertise my recent findings:

I found some results from the correlation of (previous_month;next_month) pairs. In other words, how the percentage gain in the previous month correlates with the gain of the coming month.

1. The linear trend was with positive slope, so in general it goes up the more, the more it has already gone up mo/mo.  Grin
y = 0,2394x + 0,0672
R² = 0,05661

2. If it has crashed -22% or more in the previous month, the next month has never been better than +76%.

3. If it has crashed -48% or more, the next month has also always been negative!

4. If it has gone up at least +6% and not more than +255% (LOL) in the previous month, the next month has never been worse than -32%.

5. About zero gain last month can lead to 500% gain next month (or -50% loss).


2014-5-19: Our Last Month has been -0.083 == -17%


Currently with prices floating around $500, BTC prices are about 55% down from the ATH of $1150-ish (over 4.5 months ago)... and about zero gain in the last month.  So your findings are telling us we will NOT go any lower than $250, but NO higher than $2500 within the next 30 days?   or do we measure from $340 to say we will go no higher than $1700  and no lower than $170.  

Something seems very incomplete with this supposed mathematical look at past performance to describe potential future performance.   - like it's missing something.

The important findings in my opinion were:

- If bitcoin has tanked heavily in the previous month (-22% or more), don't buy, it's a knife
- If it has gained 6%-255%, buy, because it has never been destroyed afterwards ("only" lost 32% max)

Now it is impossible to say what will happen, based on historical results, because the variance is too high.

There is a probability table prediction competition going on in the same thread.

This is a very low R^2. You are predicting 5.7% of the outcome (possibly because bitcoin's walk is primarily random).  So what this means is that all of your predictions will have huge error ranges.  With a large error range you might predict a 10% gain on average, but actually it will range from -30% to 40% and may have only a 60% chance of being greater than zero.
legendary
Activity: 2338
Merit: 1035
Risto, Do you think we will visit the 430-450 area any?

The last reversal last summer there was a pullback from 100 to 78, same patterns and everything
legendary
Activity: 2380
Merit: 1823
1CBuddyxy4FerT3hzMmi1Jz48ESzRw1ZzZ
full member
Activity: 166
Merit: 100
donator
Activity: 1722
Merit: 1036
I wish someone would take a few months to gather all the predictions made in this forum and make a list of the best nd worst predictors

You need to have a standard how to measure the goodness of the prediction.

I said that the price would go to 450 when it was 850, and it did, 2 days after. Many thought that was impressive.

If I now say that price will go to 450, it is not very impressive.

In my thread there is a prediction contest, and the idea is that the average of other predictions become the standard that you have to beat, to get some glory and fame.
legendary
Activity: 1918
Merit: 1018
I wish someone would take a few months to gather all the predictions made in this forum and make a list of the best nd worst predictors
donator
Activity: 1722
Merit: 1036
Shameless cross-post to advertise my recent findings:

I found some results from the correlation of (previous_month;next_month) pairs. In other words, how the percentage gain in the previous month correlates with the gain of the coming month.

1. The linear trend was with positive slope, so in general it goes up the more, the more it has already gone up mo/mo.  Grin
y = 0,2394x + 0,0672
R² = 0,05661

2. If it has crashed -22% or more in the previous month, the next month has never been better than +76%.

3. If it has crashed -48% or more, the next month has also always been negative!

4. If it has gone up at least +6% and not more than +255% (LOL) in the previous month, the next month has never been worse than -32%.

5. About zero gain last month can lead to 500% gain next month (or -50% loss).


2014-5-19: Our Last Month has been -0.083 == -17%


Currently with prices floating around $500, BTC prices are about 55% down from the ATH of $1150-ish (over 4.5 months ago)... and about zero gain in the last month.  So your findings are telling us we will NOT go any lower than $250, but NO higher than $2500 within the next 30 days?   or do we measure from $340 to say we will go no higher than $1700  and no lower than $170. 

Something seems very incomplete with this supposed mathematical look at past performance to describe potential future performance.   - like it's missing something.

The important findings in my opinion were:

- If bitcoin has tanked heavily in the previous month (-22% or more), don't buy, it's a knife
- If it has gained 6%-255%, buy, because it has never been destroyed afterwards ("only" lost 32% max)

Now it is impossible to say what will happen, based on historical results, because the variance is too high.

There is a probability table prediction competition going on in the same thread.
legendary
Activity: 3962
Merit: 11519
Self-Custody is a right. Say no to"Non-custodial"
Shameless cross-post to advertise my recent findings:

I found some results from the correlation of (previous_month;next_month) pairs. In other words, how the percentage gain in the previous month correlates with the gain of the coming month.

1. The linear trend was with positive slope, so in general it goes up the more, the more it has already gone up mo/mo.  Grin
y = 0,2394x + 0,0672
R² = 0,05661

2. If it has crashed -22% or more in the previous month, the next month has never been better than +76%.

3. If it has crashed -48% or more, the next month has also always been negative!

4. If it has gone up at least +6% and not more than +255% (LOL) in the previous month, the next month has never been worse than -32%.

5. About zero gain last month can lead to 500% gain next month (or -50% loss).


2014-5-19: Our Last Month has been -0.083 == -17%


Currently with prices floating around $500, BTC prices are about 55% down from the ATH of $1150-ish (over 4.5 months ago)... and about zero gain in the last month.  So your findings are telling us we will NOT go any lower than $250, but NO higher than $2500 within the next 30 days?   or do we measure from $340 to say we will go no higher than $1700  and no lower than $170. 

Something seems very incomplete with this supposed mathematical look at past performance to describe potential future performance.   - like it's missing something.


legendary
Activity: 2338
Merit: 2106
Can't help noting the price is hovering around the bullish high of Adam's poll...

EDIT: 20th April - in what time zone, Adam?

i'll go with the price of bitcoin on April 20th at 4:20PM UTC

At this time The price will be...


H I G H
legendary
Activity: 2380
Merit: 1823
1CBuddyxy4FerT3hzMmi1Jz48ESzRw1ZzZ
full member
Activity: 175
Merit: 100

 I think we're gonna go up from here by sunday mid-evening GMT at the latest. I should close my shorts soon.

 But.. just can't bring myself to do it yet.. It's dangling over this small cliff.. to ~482, or even a stairway to ~466.. I whant it.. ahh =/
donator
Activity: 1722
Merit: 1036
Shameless cross-post to advertise my recent findings:

I found some results from the correlation of (previous_month;next_month) pairs. In other words, how the percentage gain in the previous month correlates with the gain of the coming month.

1. The linear trend was with positive slope, so in general it goes up the more, the more it has already gone up mo/mo.  Grin
y = 0,2394x + 0,0672
R² = 0,05661

2. If it has crashed -22% or more in the previous month, the next month has never been better than +76%.

3. If it has crashed -48% or more, the next month has also always been negative!

4. If it has gone up at least +6% and not more than +255% (LOL) in the previous month, the next month has never been worse than -32%.

5. About zero gain last month can lead to 500% gain next month (or -50% loss).


2014-5-19: Our Last Month has been -0.083 == -17%
legendary
Activity: 2156
Merit: 1393
You lead and I'll watch you walk away.
Well, I'm all in again. Just bought with the last of my disposable fiat into Bitcoin. I don't see it going any lower for a long time, maybe never.

Welcome back to the collective. As they say in Ireland:
May the road rise up to meet you.
May the wind always be at your back.
May the sun shine warm upon your face,
and rains fall soft upon your fields.
And until we meet again,
May you hold Bitcoin in the palm of your hand.


Thanks and funny you should say that last part because I just printed my paper wallet. I'm holding all my coins in the palm of my hand. lol
hero member
Activity: 560
Merit: 500
Well, I'm all in again. Just bought with the last of my disposable fiat into Bitcoin. I don't see it going any lower for a long time, maybe never.

Welcome back to the collective. As they say in Ireland:
May the road rise up to meet you.
May the wind always be at your back.
May the sun shine warm upon your face,
and rains fall soft upon your fields.
And until we meet again,
May you hold Bitcoin in the palm of your hand.
legendary
Activity: 2156
Merit: 1393
You lead and I'll watch you walk away.
Well, I'm all in again. Just bought with the last of my disposable fiat into Bitcoin. I don't see it going any lower for a long time, maybe never.
Jump to: