I have been fervently upgrading my methodology regarding this exponential trendline issue. Currently the model which I personally use contains the following data:
- Before Mt.Gox, fixed 0.005
- After Mt.Gox until 31.12.2012, Mt.Gox daily volume weighted average
- 1.1.2013 and later, VWA of Mt.Gox, Bitstamp and BTC China daily VWA's.
The trendline is recalculated monthly back in time (to reflect the trendline available in any point of time in history) and will be recalculated biweekly in the future. The model did not seem to be least squares but rather simple fit (equal area above and below trendline). The slope of the trend is established very well during the last year - the recalculations of 2013 have seen the slope fluctuating in the band of only 2.9%, and the price has moved from below to above, to below, to above, and again down towards the trendline.
Others use Mt.Gox data and ignore the history before 17.7.2010. This results in a trend which is 15% less steep in logarithmic scale, and shows the current price ($600) as even more overvalued than my model.
This model has given insight on the buyback level after the crash has played itself out. As such, it would be foolhardy to assume that the current crash would play itself out exactly as the April one did. But the trendline analysis seems to confirm the essential similarities to such a degree that we can make an informed guess on how low to go and when. It seems inevitable that we will touch the trendline soon on a daily basis (in the recent dip it was only crossed in BTC China and Bitstamp, and not on a daily VWA). After touching it, we will go below it. Last summer we spent 158 days at an average of -0.222 log units (-40%) below the trendline. This period started 55 days after the top.
Today the trendline is at $446 and growth is $3.15 daily. There are many ways to try to estimate how quickly we will go to -40% below trendline. One (obviously arbitrary) method is to take the final capitulation of the last bubble, and compare its duration from the top and price level relative to trendline. That would suggest a $339 capitulation in 26.2.2014.
The last bubble saw 3 distinct buying opportunities, 2 of which happened relatively quickly after the top as dips/flashcrashes, and the 3rd one was the final capitulation much after that lasted for days. Now we have only seen one flashcrash that has taken us to long-term buy levels, so at least one more is expected. The reason for the expectation is not that there is some cosmic magic that the two bubbles should play out similarly (afaik there is not) but that the trendline, which has taken into account the exchange rate over the course of Bitcoin's history (and thus is the sum of every factor ever that has, could have or will have affected the price) suggests we are overvalued, and the downtrend will quite likely end only when this overvaluation is purged. This is deeply rooted in experience from other markets over centuries. Don't jump under the train that's falling from the sky.
If this thread is to help us in speculation, the advice it would give is as follows:
- Do not buy back yet in $500, for that would be catching the falling knife. Even if it bounces back from above $500, it will come down again. Last summer the final one was weeks after the interest had started to wane. When my usual critics don't care to argue any more but have thrown in the towel and don't care anymore, then we will have cheap coins. It is more likely than not that $3XX will come.
