rpietila Wall Observer - the Quality TA Thread ;)

[bitfair]

Predicting a range is clearly useful, but having the value fall into that range or hit the mean value of the range does not always mean the prediction is better.

Example:
Prediction 1: 90-100; Prediction 2: 50-150. Actual: 101
The first is clearly the better prediction, although the actual is outside the range and further from the mean than the second.

What you really want to do is 1) reward a predicted mean value (i.e. max+min/2) that is close to the actual; 2) reward a smaller range preferentially to a larger range.

Actually, it is not clear that the first prediction is better. It depends. Also, to the point 1), consider that the probability density function can be asymmetric within the range, i.e. the mean can be different from 0.5*(max+min).

No. D is the deviation from the interval edge and is added to the weighted interval width. I think you misread it.

Sorry, I misread your AND as an OR in the conditional and I didn't look too closely at the inequalities.

In that case, this proposal would give the same score to two forecasts with the same I, regardless of where within the interval the actual outcome falls. For example: a prediction of 90-110 would be ranked equal to a prediction of 99-119 with an outcome of 100. I don't find that intuitive either.

[bb113]

Predicting a range is clearly useful, but having the value fall into that range or hit the mean value of the range does not always mean the prediction is better.

Example:
Prediction 1: 90-100; Prediction 2: 50-150. Actual: 101
The first is clearly the better prediction, although the actual is outside the range and further from the mean than the second.

What you really want to do is 1) reward a predicted mean value (i.e. max+min/2) that is close to the actual; 2) reward a smaller range preferentially to a larger range.

Actually, it is not clear that the first prediction is better. It depends. Also, to the point 1), consider that the probability density function can be asymmetric within the range, i.e. the mean can be different from 0.5*(max+min).

No. D is the deviation from the interval edge and is added to the weighted interval width. I think you misread it.

Sorry, I misread your AND as an OR in the conditional and I didn't look too closely at the inequalities.

In that case, this proposal would give the same score to two forecasts with the same I, regardless of where within the interval the actual outcome falls. For example: a prediction of 90-110 would be ranked equal to a prediction of 99-119 with an outcome of 100. I don't find that intuitive either.

Ok in that case people will have to guess distributions rather than intervals, then just compare the probability density at the price. If people want to do intervals, then they can use uniform distributions.

[envy2010]

Actually, it is not clear that the first prediction is better. It depends. Also, to the point 1), consider that the probability density function can be asymmetric within the range, i.e. the mean can be different from 0.5*(max+min).

Yes, it depends on how you define "better". I'd consider the first prediction more useful and therefore better, even though the actual falls outside the range. That's what range is not a good method.

A more precise method would be to assume a normal distribution on a log scale and have the prediction include both mean and standard deviation so one could compute the probability associated with a given price. The prediction with the high probability  at the final actual value wins.

[bitfair]

Ok in that case people will have to guess distributions rather than intervals, then just compare the probability density at the price. If people want to do intervals, then they can use uniform distributions.

This sounds like a solid, sensible and fair solution.

Loved the illustrations too!

Edit: May be a bit cumbersome for each forecaster, but allows each forecaster to put as much work into it as he wants and still remains fair.

[pa]

Why not have an exclusive channel to publish analysts' predictions? After a certain delay-period, all predictions would become public. The market price for timely access to an analyst's predictions would be a good metric of the accuracy of that analyst.

[bb113]

I realized that the sum of all the guesses would give a fancy consensus prior on the price at a given date too...

[SlipperySlope]

I realized that the sum of all the guesses would give a fancy consensus prior on the price at a given date too...

The Wisdom of Crowds.

[bb113]

I realized that the sum of all the guesses would give a fancy consensus prior on the price at a given date too...

The Wisdom of Crowds.

Perhaps it works by averaging the weights of all predictors to near 1?
http://heatherlench.com/wp-content/uploads/2008/07/dawes2.pdf

[thezerg]

u want to generate the gaussian who's two points that are 1 std dev from the center fall on the 2 predicted points.  normalize the curve so Y is 1 at those pts.  Plug in the actual BTC log value and then divide by the width of the prediction to favor narrow range predictions.  highest number wins.

[Siegfried]

I liked when this thread was about Bitcoin analysis, not analyzing how to analyze whose analysis is best.

[bb113]

I liked when this thread was about Bitcoin analysis, not analyzing how to analyze whose analysis is best.

But why would you trust a bitcoin analysis that hasn't been analyzed? Etc, until first principles. Greeks dude, the greeks got that far.

[SlipperySlope]

I liked when this thread was about Bitcoin analysis, not analyzing how to analyze whose analysis is best.

Actually in recent days, there has been much more emphasis on charts and such than was the case before.

In my opinion,the main issue to be resolved was - is the bottom of the November 2013 bubble behind us, or is another leg downward ahead of us? The preponderance of technical evidence debated here - is that the bottom was at $339 on April 10 using prices reported by Bitstamp.

[Siegfried]

I was being mildly facetious. We will all appreciate a solid formula for evaluating the accuracy of analysis.

[Trolololo]

I have updated the fractal (Mtgox 2010-2011 vs Bitstamp 2013-2014 superbubbles shape comparison chart) on 1d bars.

Strechting timescale to fit 0.50-1.10 vs 259-1163 maximums
and strechting pricescale to fit shapes,

the shape and timeframe of maxs and mins comparison is amazing:

[Trolololo]

Here is the full fractal (1d bars):



Although I think that the probability of this fractal is quite small (but growing every day), it's interesting to visualize the speed of 2011 superbubble in today's prices:

- over 700 by end of April, clearly breaking de 2014 downtrend (10-15 days from now)
- ATH by mid May 2014 (1 month from now)
- around 10,000 by mid June 2014 (2 months from now)
- around 50,000 by mid July 2014 (3 months from now) 100x in just 3 months from now!!!
- around 120,000 by mid Sept 2014 (5 months from now)
- around 5,000 by Oct 2015 (1 year of heavy correction)

If we are over 700 by end of month...

[MahaRamana]

The chances of this happening are small.
Most bitcoin holders would not be able to hold through a 100X or 200X increase happening so fast.
Bitcoin holders would find themselves sitting on trillions of wealth forcing them to diversify away from bitcoin. This would prevent such steep and sustained increase.
The only scenario allowing this kind of price increase in 2014 is a hyper inflationary collapse of the USD along with a global meltdown of the financial and monetary system. Such scenario does not have a zero probability at all, but I would not want it to happen.

A nuclear war between the US/EU and Russia/China could make it happen as well - but it also guarantees a hyper inflationary collapse.

[conspirosphere.tk]

If we are over 700 by end of month...

now it's more interesting to figure out how low it can go

[TERA]

Here is the full fractal (1d bars):



Although I think that the probability of this fractal is quite small (but growing every day), it's interesting to visualize the speed of 2011 superbubble in today's prices:

- over 700 by end of April, clearly breaking de 2014 downtrend (10-15 days from now)
- ATH by mid May 2014 (1 month from now)
- around 10,000 by mid June 2014 (2 months from now)
- around 50,000 by mid July 2014 (3 months from now) 100x in just 3 months from now!!!
- around 120,000 by mid Sept 2014 (5 months from now)
- around 5,000 by Oct 2015 (1 year of heavy correction)

If we are over 700 by end of month...

This is legit TA. $250,000 by September confirmed.

[wachtwoord]

SNIP

This is legit TA. $250,000 by September confirmed.

Come on, it's a fun thought experiment at least.

There's no such thing as legit TA

[MatTheCat]

This is legit TA. $250,000 by September confirmed.

I will take the notional other side of this legit TA and say $250 dollars by mid/late May 2014.