Topic: Has Bitcoin Gone for a Random Walk? (Read 1232 times)
2012 results: highly significant lagged variables, therefore NOT A RANDOM WALK.
. reg dt0 dt1 dt2 t3 if(time<365)
Source | SS df MS Number of obs = 364
-------------+------------------------------ F( 3, 360) = 7.55
Model | .058269935 3 .019423312 Prob > F = 0.0001
Residual | .926368985 360 .002573247 R-squared = 0.0592
-------------+------------------------------ Adj R-squared = 0.0513
Total | .98463892 363 .002712504 Root MSE = .05073
------------------------------------------------------------------------------
dt0 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dt1 | -.2417949 .0523137 -4.62 0.000 -.3446738 -.1389161
dt2 | -.1060391 .0522126 -2.03 0.043 -.2087191 -.0033591
t3 | -.0045321 .0069005 -0.66 0.512 -.0181024 .0090381
_cons | .0126326 .0142685 0.89 0.377 -.0154274 .0406926
------------------------------------------------------------------------------
2013 results: no statistically significant lagged variables, therefore we cannot reject random walk.
. reg dt0 dt1 dt2 t3 if(time>365 & time<730)
Source | SS df MS Number of obs = 364
-------------+------------------------------ F( 3, 360) = 0.34
Model | .006482693 3 .002160898 Prob > F = 0.7979
Residual | 2.3014901 360 .006393028 R-squared = 0.0028
-------------+------------------------------ Adj R-squared = -0.0055
Total | 2.30797279 363 .006358052 Root MSE = .07996
------------------------------------------------------------------------------
dt0 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dt1 | .0132722 .0527079 0.25 0.801 -.0903819 .1169262
dt2 | -.0135911 .052703 -0.26 0.797 -.1172356 .0900534
t3 | -.003767 .0040118 -0.94 0.348 -.0116564 .0041224
_cons | .0285126 .0191565 1.49 0.138 -.00916 .0661852
------------------------------------------------------------------------------
2014 result: Significant variable to 99% confidence level, so can reject random walk.
. reg dt0 dt1 dt2 t3 if(time>730)
Source | SS df MS Number of obs = 363
-------------+------------------------------ F( 3, 359) = 2.84
Model | .013178025 3 .004392675 Prob > F = 0.0380
Residual | .555674788 359 .001547841 R-squared = 0.0232
-------------+------------------------------ Adj R-squared = 0.0150
Total | .568852812 362 .001571417 Root MSE = .03934
------------------------------------------------------------------------------
dt0 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dt1 | -.0676219 .0522455 -1.29 0.196 -.1703676 .0351238
dt2 | -.1358541 .0522824 -2.60 0.010 -.2386725 -.0330358 t3 | -.0070585 .0077177 -0.91 0.361 -.0222361 .008119
_cons | .041105 .0481442 0.85 0.394 -.0535751 .1357851
------------------------------------------------------------------------------
So I did it for you. I tested whether the daily price series was a random walk or not in the past. I got the opposite results from you.
Firstly, price definitely WAS NOT a random walk in 2012. Totally not random.
Conversely, I cannot reject the random walk hypothesis for 2013. It actually seems quite random that year, at least until end November.
I can reject random walk in 2014. The persistent bear market in 2014 IS NOT A RANDOM WALK.
This is simple statistical testing. I'm surprised you didn't do it before presenting your analysis.
Firstly, price definitely WAS NOT a random walk in 2012. Totally not random.
Conversely, I cannot reject the random walk hypothesis for 2013. It actually seems quite random that year, at least until end November.
I can reject random walk in 2014. The persistent bear market in 2014 IS NOT A RANDOM WALK.
This is simple statistical testing. I'm surprised you didn't do it before presenting your analysis.
For all the fancy charts, you don't seem to have done the simplest thing of all.
Have you not statistically tested whether the price series IS or IS NOT a random walk? Is the change in the price today correlated with the price yesterday (or any previous day), or not?
All I need are the co-efficients of the lagged price variables and their standard errors. I can do the rest...
http://en.wikipedia.org/wiki/Random_walk_hypothesis
Thanks
Have you not statistically tested whether the price series IS or IS NOT a random walk? Is the change in the price today correlated with the price yesterday (or any previous day), or not?
All I need are the co-efficients of the lagged price variables and their standard errors. I can do the rest...
http://en.wikipedia.org/wiki/Random_walk_hypothesis
Thanks
Thanks for the analysis. Now that I have accumulated a bit more at this low I will resume my holding strategy.
More like a drunk walk.
Definitely intriguing from a trading perspective, so now that the rwi of the blue is approaching 3.0 a short should be opened but knowing when to close it as it's not far away from 3 is the tricky part for a newcomer.
like my dog. That goes for 'random walks' too.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
Cool story, bro!!!
Now, go and learn proper English.
like my dog. That goes for 'random walks' too.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
Cool input, bro
like my dog. That goes for 'random walks' too.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
No man, it has failed already. There will be no massadoption anytime soon. It's overpriced.
Has Bitcoin Gone for a Random Walk?
This analysis will look at bitcoin’s cycles of volatility and stability to identify trading strategies during random walks.
Background Information
What is Random Walk Theory?
A random walk is a large pattern that is formed by the cumulative effect of small changes. This is often applied in finance to mean the price of an asset ‘drifting’ in a particularly direction, with no major price swing being the sole cause. Consequently, price changes appear to be random and hard to predict.
What do random walks have to do with bitcoin?
Although bitcoin is (in)famous for its volatility, the price has been through two periods of random walks. These periods are the majority of 2012 and the second half of 2014.
Why should bitcoin traders care about random walks?
As our analysis shows, traders should adopt different trading strategies when bitcoin enters phases of random walks. Namely, keeping positions open for longer and using more leverage can help traders earn returns during random walks. This is because price changes are less sudden and smaller, requiring greater patience and more funds to achieve the same profits enjoyed outside of random walks.
Bitcoin Technical Analysis
Introduction to analysis
This analysis was made using a Random Walk Index (RWI) and the Chart Mill Value indicator (CVI).
The RWI measures the strength of market movements by comparing price movements to acceptable trading ranges. Thus a small price movement can be explained by random walk, while larger movements are part of a larger market trend.
The CVI identifies price ranges that are over-bought or over-sold. Relying on moving averages to identify extreme price deviations is less accurate because a long-term price deviation is hidden when the moving average follows the price closely. This happens frequently during random walks. In contrast, the CVI divides the spread by the average true range to give more accurate results.
The TradingView scripts were written by LazyBear. This author highly recommends following LazyBear on TradingView and thanks him for his scripts that made this analysis possible. LazyBear’s extensive range of scripts can be downloaded here.
Long-term bitcoin analysis
One should read this chart from the bottom to the top.
RWI chart
Starting from the bottom of the chart, the RWI is plotted with a red line for price highs, and a blue line for price lows. When these two lines are closer together, the price is less volatile — thus, price movements are much more likely to be a result of random of walk. Conversely, when the two lines are further away, price changes are more likely to be driven by market trends, particularly greed-fear cycles.
When the RWI of price highs (red line) is above 1, there is a good chance of a sustainable rally. However when the RWI of price lows (blue line) is above 1, there is a good chance of a deep correction. A trader can use these indicators in the same way that moving averages are used to identify bullish or bearish trends. The histogram above this simply presents the same data on just one scale.
A key observation from this RWI chart is that crossovers appear to be consistent indicators of an incoming change in price direction.
It looks like the next crossover will occur in late March 2015, which is shown by the RWI chart; this is consistent with support/resistance levels.
Logarithmic chart
It is apparent that bitcoin enters phases of bubbles, separated by random walks. This becomes more clearer when charting basic long-term support and resistance levels.
With a good degree of fit, it looks as if there is a channel in which bitcoin enters a random walk. When the price drifts out of this random walk channel, it is because of a significant market trend. The price breaking out of this channel is a very bullish sign, with further gains to be made in most occasions.
This random walk channel theory provides good evidence for bitcoin’s price following a biennial (two year) cycle: a volatile year, and then a year of random walk. If one were to add data from Mt. Gox, a biennial cycle is even clearer:
2011 — Price rally
2012 — Positive random walk
2013 — Two price rallies
2014 — Negative random walk
If this pattern holds true, one can expect 2015 to feature a price rally, followed by a positive random walk in 2016.
Let’s take a look at how this information can be used for short-term bitcoin trading.
Short-term bitcoin analysis
This chart should be read from the top to the bottom.
Logarithmic chart
The top price chart simply examines the price directly before and after entering the random walk channel. The trend line being broken in July is when bitcoin starts the random walk.
By simply looking at the chart, the price exhibits more gradual price movements, which are indicative of a random walk, after July. Up to July, the price is much more volatile.
RWI chart
This time the RWI chart is using seven periods (days in this case) to assess short-term volatility. It is interesting to see that there are fewer large price movements — defined as movements greater than a RWI of 4 — after the price starts its random walk. Before July, the price mostly changes direction at a RWI of 4. After July, volatility rarely exceeds a RWI of 3.
A trader can use this RWI chart to time the opening and closing of short positions. Namely, when the RWI of price lows (blue line) is increasing and moving towards a RWI of 3, a short position should be opened. When a RWI of 3 is reached, the position should be closed. This is the acceptable range of price movement within a random walk, and the market is likely to start correcting at this point.
CVI chart
The CVI chart identifies over-bought/sold prices with greater accuracy than moving averages. Unfortunately, since bitcoin entered the random walk, there is less scope to profit from a less volatile price. This means that in percentage terms, traders are likely to be experiencing lower returns since July.
However, traders can still achieve the same returns in BTC. If leverage / margin trading is used, the same profits in BTC can be made because the trader is using more funds to compensate for the lack of volatility. If the volatility drops by 50%, trading with 100% larger positions will lead to equal gains of BTC.
Furthermore, beginner traders could benefit from using a simple CVI chart to time entry and exit positions. By trading to this chart, emotions can be removed from trading decisions, leading to buying when the price is low and selling when the price is high. This method helped BTC.sx CEO turn $100 into $200,000.
Conclusion
The purpose of this analysis has been two-fold: one, analyse whether bitcoin is following a random walk; two, provide trading recommendations. The key trading recommendations can be summarized as follows:
1. Watching the long-term RWI chart for crossovers can help traders find the start of bullish or bearish price trends.
2. Significant levels of price volatility is not expected until late March 2015.
3. When the RWI of price lows (blue line) is increasing and moving towards a RWI of 3, a short position should be opened. When a RWI of 3 is reached, the position should be closed.
4. The CVI chart can be used to time the opening and closing of positions with greater accuracy than moving averages.
5. To compensate for the lack of volatility, consider trading with leverage.
Written by Josh Blatchford, CMO of BTC.sx, a bitcoin trading platform that supports up to 10x leverage on Bitfinex, Bitstamp and itBit
This analysis will look at bitcoin’s cycles of volatility and stability to identify trading strategies during random walks.
Background Information
What is Random Walk Theory?
A random walk is a large pattern that is formed by the cumulative effect of small changes. This is often applied in finance to mean the price of an asset ‘drifting’ in a particularly direction, with no major price swing being the sole cause. Consequently, price changes appear to be random and hard to predict.
What do random walks have to do with bitcoin?
Although bitcoin is (in)famous for its volatility, the price has been through two periods of random walks. These periods are the majority of 2012 and the second half of 2014.
Why should bitcoin traders care about random walks?
As our analysis shows, traders should adopt different trading strategies when bitcoin enters phases of random walks. Namely, keeping positions open for longer and using more leverage can help traders earn returns during random walks. This is because price changes are less sudden and smaller, requiring greater patience and more funds to achieve the same profits enjoyed outside of random walks.
Bitcoin Technical Analysis
Introduction to analysis
This analysis was made using a Random Walk Index (RWI) and the Chart Mill Value indicator (CVI).
The RWI measures the strength of market movements by comparing price movements to acceptable trading ranges. Thus a small price movement can be explained by random walk, while larger movements are part of a larger market trend.
The CVI identifies price ranges that are over-bought or over-sold. Relying on moving averages to identify extreme price deviations is less accurate because a long-term price deviation is hidden when the moving average follows the price closely. This happens frequently during random walks. In contrast, the CVI divides the spread by the average true range to give more accurate results.
The TradingView scripts were written by LazyBear. This author highly recommends following LazyBear on TradingView and thanks him for his scripts that made this analysis possible. LazyBear’s extensive range of scripts can be downloaded here.
Long-term bitcoin analysis
One should read this chart from the bottom to the top.
RWI chart
Starting from the bottom of the chart, the RWI is plotted with a red line for price highs, and a blue line for price lows. When these two lines are closer together, the price is less volatile — thus, price movements are much more likely to be a result of random of walk. Conversely, when the two lines are further away, price changes are more likely to be driven by market trends, particularly greed-fear cycles.
When the RWI of price highs (red line) is above 1, there is a good chance of a sustainable rally. However when the RWI of price lows (blue line) is above 1, there is a good chance of a deep correction. A trader can use these indicators in the same way that moving averages are used to identify bullish or bearish trends. The histogram above this simply presents the same data on just one scale.
A key observation from this RWI chart is that crossovers appear to be consistent indicators of an incoming change in price direction.
It looks like the next crossover will occur in late March 2015, which is shown by the RWI chart; this is consistent with support/resistance levels.
Logarithmic chart
It is apparent that bitcoin enters phases of bubbles, separated by random walks. This becomes more clearer when charting basic long-term support and resistance levels.
With a good degree of fit, it looks as if there is a channel in which bitcoin enters a random walk. When the price drifts out of this random walk channel, it is because of a significant market trend. The price breaking out of this channel is a very bullish sign, with further gains to be made in most occasions.
This random walk channel theory provides good evidence for bitcoin’s price following a biennial (two year) cycle: a volatile year, and then a year of random walk. If one were to add data from Mt. Gox, a biennial cycle is even clearer:
2011 — Price rally
2012 — Positive random walk
2013 — Two price rallies
2014 — Negative random walk
If this pattern holds true, one can expect 2015 to feature a price rally, followed by a positive random walk in 2016.
Let’s take a look at how this information can be used for short-term bitcoin trading.
Short-term bitcoin analysis
This chart should be read from the top to the bottom.
Logarithmic chart
The top price chart simply examines the price directly before and after entering the random walk channel. The trend line being broken in July is when bitcoin starts the random walk.
By simply looking at the chart, the price exhibits more gradual price movements, which are indicative of a random walk, after July. Up to July, the price is much more volatile.
RWI chart
This time the RWI chart is using seven periods (days in this case) to assess short-term volatility. It is interesting to see that there are fewer large price movements — defined as movements greater than a RWI of 4 — after the price starts its random walk. Before July, the price mostly changes direction at a RWI of 4. After July, volatility rarely exceeds a RWI of 3.
A trader can use this RWI chart to time the opening and closing of short positions. Namely, when the RWI of price lows (blue line) is increasing and moving towards a RWI of 3, a short position should be opened. When a RWI of 3 is reached, the position should be closed. This is the acceptable range of price movement within a random walk, and the market is likely to start correcting at this point.
CVI chart
The CVI chart identifies over-bought/sold prices with greater accuracy than moving averages. Unfortunately, since bitcoin entered the random walk, there is less scope to profit from a less volatile price. This means that in percentage terms, traders are likely to be experiencing lower returns since July.
However, traders can still achieve the same returns in BTC. If leverage / margin trading is used, the same profits in BTC can be made because the trader is using more funds to compensate for the lack of volatility. If the volatility drops by 50%, trading with 100% larger positions will lead to equal gains of BTC.
Furthermore, beginner traders could benefit from using a simple CVI chart to time entry and exit positions. By trading to this chart, emotions can be removed from trading decisions, leading to buying when the price is low and selling when the price is high. This method helped BTC.sx CEO turn $100 into $200,000.
Conclusion
The purpose of this analysis has been two-fold: one, analyse whether bitcoin is following a random walk; two, provide trading recommendations. The key trading recommendations can be summarized as follows:
1. Watching the long-term RWI chart for crossovers can help traders find the start of bullish or bearish price trends.
2. Significant levels of price volatility is not expected until late March 2015.
3. When the RWI of price lows (blue line) is increasing and moving towards a RWI of 3, a short position should be opened. When a RWI of 3 is reached, the position should be closed.
4. The CVI chart can be used to time the opening and closing of positions with greater accuracy than moving averages.
5. To compensate for the lack of volatility, consider trading with leverage.
Written by Josh Blatchford, CMO of BTC.sx, a bitcoin trading platform that supports up to 10x leverage on Bitfinex, Bitstamp and itBit
Jump to: