This series deals with the question if the crypto markets follow the random-walk theory, or if the markets have a favourable side. More specifically we look at the data to find out, if there is a favourable side of the market i.e. if it is favourable to be on the long/short side for different cryptocurrencies. The analysis focuses on daily price data and we just look at the relation of up-/down-close candles. The cryptocurrency I look at in this article is Bitcoin (BTC), but I'll extend my analysis to other major crytocurrencies in the future. Be wary to take this analysis with a grain of salt, as the data available might not be enough to extract actual valid information from the data. Further this must not be seen as financial advice.
To get the price data I utilized the CCXT library to query price data from the Gate.io Exchange API. The reason for using Gate.io is, that it provides price data for BTC/USDT all the way back to April of 2013. Amongst the exchanges supported by the CCXT library this is the farthest back price data. To get an overview of all supported exchanges in combination with their earliest price data for BTC/USDT and their request data limit you can check out this article.
Let's assume the market is random i.e. that the probability for an up-closing candle is equal to the probability for a down-closing one. If this is the case we would expect a pretty equal amount of up- and down-close candles throughout the available price data. Interestingly for BTC this is not the case. As you can see in the chart below, in the time from March 31st 2013 to August 1st 2022 (3411 Days), BTC printed more up-closing candles than down-closing candles. More specifically BTC printed 1793 up-close candles, which corresponds to 52.6% of all the candles, opposed to only 1590, or 46.6%, down-close candles. The remaining 28 days (0.8%) were days, where the opening price was equal to the closing price, i.e. Equal Days.
This difference in percentage of up- and down-close days can be seen as an indicator, that BTC does not necessarily follow the random-walk theory, which many believe to be the foundation of the financial markets. But this disparity in up- and down-close days is not enough to be able to confidently say, that each day BTC has a higher probability of going up than down, which would be against the random-walk theory. This is due to the case, that we could say, that the skew to the up-side is brought about by the fact, that BTC has been in a general uptrend from 2013 to 2022. Therefore we have to look at the data more closely to find out if Bitcoin's movement is really not that random after all.
We do this by looking at the probabilities of up-/down-close days following one and two consecutive down-close days. As is shown in the figure below BTC had an up-close day 55.1% of the time following one down-close day and it had an up-close day 55.7% of the time following two consecutive down-close days.
This goes against the random-walk theory, as if BTCs movement would really be random, we would expect the probability of price going up and down to be equal, with the previous days movement having no impact. But looking at the data we see, that BTC moves up more often than down after price moved down the previous day (or the previous two days). This supports the skew of the probabilities to the up-side, which we have seen earlier and allows the suspicion, that for BTC it might be favourable to be on the long side (no financial advice though).
To fully conclude this analysis we also have to look at what happens to the probabilities of up-/down-close days after one and two consecutive up-close days. So inversely to the previous figure, where we looked at the probabilities of BTC closing up/down after one and two down-close days, we now look at the probabilities of BTC closing up/down after one and two up-close days. Interestingly, as presented in the figure below, here the percentages point towards a really random movement i.e. BTC closes upwards equally as often as it closes downwards following one and two up-close days.
We looked at the price data of BTC from March 31st 2013 to August 1st 2022 and analyzed if the movement of BTC is truly random, or if there might be a favourable side of the market. Both, the split of Up, Down, and Equal days and the probabilities of an up-/down-close day following one and two down-close days point to a non-random movement and a skew of the probabilities to the long side. The only case where the analysis showed random movement for BTC were the probabilities of up-/down-close days following one and two up-close days.
What we can take away from this is, that when BTC moves up the direction of the next candle tends to be random and when BTC moves down the direction of the next candle is more probable to be up than down. So in general this suggests random price movement for BTC as long as the trend is upwards.