Bitcoin Forum

Hi everyone.

Does anyone have a download of historical prices and difficulties? Optimally weekly, monthly or maybe even daily in a csv or other table format?

I would like to do some statistics on it to see whether or not for example difficulty and price correlate with each other long and short term, what events have been total outliers and so on.

Any suggestions for that?

Correlation does not imply causation.

http://wickershamsconscience.files.wordpress.com/2011/12/690correlation.jpg

nice

Actually, there might be some truth to that one. Some scientists were caught faking their temperature data. Some weather stations were found out to be located near growing cities, and their measurements included a lot of "temperature noise" generated by those cities. Obviously, the more money you pour into global warming research, the more incentive there is to scare people with global warming, the more money governments pour into global warming research...

So we understand that "Correlation does not imply causation." but we still believe in technical analysis. Makes no sense to me.

"In logic, the technical use of the word 'implies' means 'to be a sufficient circumstance.' This is the meaning intended by statisticians when they say causation is not certain. Indeed, p implies q has the technical meaning of logical implication: if p then q symbolized as p → q. That is 'if circumstance p is true, then q necessarily follows.' In this sense, it is always correct to say 'Correlation does not imply causation.'

However, in casual use, the word 'imply' loosely means suggests rather than requires. The idea that correlation and causation are connected is certainly true; where there is causation, there is likely to be correlation. Indeed, correlation is used when inferring causation; the important point is that such inferences are made after correlations are confirmed to be real and all causational relationship are systematically explored using large enough data sets.

Edward Tufte, in a criticism of the brevity of 'correlation does not imply causation,' deprecates the use of 'is' to relate correlation and causation (as in 'Correlation is not causation'), citing its inaccuracy as incomplete. While it is not the case that correlation is causation, simply stating their nonequivalence omits information about their relationship. Tufte suggests that the shortest true statement that can be made about causality and correlation is one of the following:

'Empirically observed covariation is a necessary but not sufficient condition for causality.'
'Correlation is not causation but it sure is a hint.'"

- http://en.wikipedia.org/wiki/Correlation_does_not_imply_causation#Usage

Clearly you have the cause and effect of the Shyamalan one backwards. His movies started to suck before the newspapers sales dropped off. Therefore, we can blame the decline of newspapers on Shyamalan's crappy movies.

+1

Correlation does not imply causation, but absence of causation does not preclude predictive power. In fact, correlation implies predictive power.

Yes, cum hoc ergo propter hoc.

I use that sentence all the time, I do research for a living, although in a completely different setting.
We have another saying: Causation does not exist without correlation.

Of course in general statistics you can easily have outliers and problems. That is why within this analysis, you can possibly find events and their effects, and calculate them out.

You can take some timelines and see whether or not certain difficulty increases can be regressed to have caused "a certain amount of variance" explanation. You can pull a multi point longitudinal analysis.
I know that all of those are useless for prediction, however the strength of the linkage is what interests me.

You see, if I find that calculating events out, I find statistically significant correlations of above .5 with p lower .01, I might even invest in more 7970s. If the SD is too high or alpha "cleaned" variance cannot be explained to at least 40-50%, I will not.

I would not use this model for anything more than this kind of buying decision. Any other, bigger speculation would be voodoo. Even this one is pushing it ;)
 

http://www.scholarpedia.org/article/Predator-prey_model

Has anyone looked into tearing down the right edge of that mountain to get rid of murder? Someone should really look into that.

Absolute tosh. Not even worth commenting upon. Oh wait :(

Some people are caught talking bollox, but that doesn't mean that everyone is talking bollox.....

The vast majority of science now agrees that human induced climate change is fact. Your job, then, is to disprove it. That's how science works. And if they're right (and you're wrong), then doing nothing to tackle climate change is a crime against humanity. Yet if you're right (and science is wrong), then the worst that happens is we all get some rather nice wind turbines (IMHO). So I vote for doing something.

So there.

*Facepalm*

They had to adjust the temperatures because of factors you mention above.

If they did not "fake" the temperature data, it would seem like the temperature was rising a lot more than they do.
Even something simple as if a tree is allowed to grow too close to a weather-station, means that the temperatures have to be corrected. Also new building/trees further away that blocks part of the sun during the day have to be corrected. - Not to mention if something is removed giving more sun at the weather-station spot. Then there are changes in the way the wind flows due to vegetation/buildings etc.

The temperature-datas have to be corrected "all the time" because one tries to compare baseline datapoints in a dynamic world.

It's the same with sea-levels. The landmasses are rising/sinking at different speeds all over the world and there is a lot of factors that determines the ebb and flow than "Sun goes up, sun goes down/Tide goes in, tide goes out".

In Figure 2. Is global warming a hoax propagated by scientists? the correlation weakens significantly when you account for inflation. $69.8m in 1993-dollars is actually $103.63m in 2009-dollars if you calculate it with CPI. Of course science isn't using consumer goods so Science Price Inflation (SPI) could be alot higher. Even if it is not, the CPI is clearly manipulated by the government and the real rate of inflation is higher than CPI (I didn't want to pay for a subscription to use the shadow stats (http://www.shadowstats.com/inflation_calculator?amount1=69.8&y1=1993&m1=1&y2=2009&m2=1&calc=Find+Out) calculator so I used a ruler on my screen to compare the bars and found that according to the site, $69.8m 1993-dollars was about $258m 2009-dollars, hence the National Science Foundation's R&D budget has actually been reduced).

On the other hand, accounting for inflation gives Figure 4. Would M. Night Shyamalan start making good movies again if people bought more newspapers a stronger correlation...hold on, did Shyamalan ever make any good movies?!

So, anyone has basically harped on the idea "difficulty vs. price."

I have historical price data now from Gox, anyone has accurate historical difficulty data?


Or would I actually need to pull that off the blockchain by hand?

http://upload.wikimedia.org/wikipedia/commons/a/aa/Volterra_lotka_dynamics.PNG

truth, but reality is a tad more complicated than that.

Non periodic solution looks like that:

http://i140.photobucket.com/albums/r6/kukulaj/ResourcePPR.jpg

This is called an evolutionary stable strategy.

I think people are to harsh on OP.

Although there may not be a definitive correlation between hashing power/difficulty, it surely does say something about the interest/commitment to BTC.

I for one would like to see a chart like this. Ideally it should be corrected for advances in computing power/cost.

With minimal value people would still mine, but when a lot of people invest in fairly expensive dedicated hardware etc. just to mine, it shows a dedication and confidence in BTC value.

Fun thing is, people love to rely in these things in trading, but Nash them Art other times.

Statistics are merely tools to depict certain constructs in mathematical language.

There are some ideas I have had to analyse this. Psychologists use the constructs I am thinking of with quite success, although in this regard we will never reach the accuracy of physics.

What I want, actually, is to find connections between variables and their impact on price. I would try to get enough significant data points and then create a model that shows data points that do not conform to the model. Finding more and more points with actual probable causes, I would then try to run a tractor analysis, considering the data is sufficient.

This way, I would gain some understanding in short, mid any long term variance of how it has played out so far. This goes only so far in prediction value, but it could give me some insights into how the whole trading community thinks on a systems models.

I would then write a paper about it and try to get to t published in a journal on behavioural economics.

Whether or not it will have predictive value is up in the stars. It could use news indicators and weight them like other trading systems, but it could be pure voodoo as well.

Just because things are correlated doesn't mean they aren't causing each other.

Sometimes when two things appear to be linked, they really are.  Throwing out a bunch of examples of things that aren't isn't helpful at all.

Anyone who has thought about mining should realize there is a very clear and obvious reason why the two are linked.  The first thing I do before buying mining hardware is attempt to estimate the time to payoff.  If it's longer than 6 month, it's probably not worth buying the hardware.  So, if the difficulty drops in relation to price, it makes a lot more sense to buy hardware and mine, thus pushing difficulty up.  If price drops while difficulty remains high, or if difficulty increases and price does not, it makes more sense to BUY bitcoin outright and forget about mining.  As a result, the value of bitcoin and cost of mining (directly associated with difficulty) correlate fairly well over the long term.

And the this would be typical calcs I would want to do... Xx70 prices vs. Price, FPGA price vs price, difficulty vs. Price.

And then search for data points that do NOT fit such models. Fun stuff for the free time of a novice researcher, if you ask me.