Up until now, Project-Doge has been focused on training an AI to buy and sell dogecoin (and more recently bitcoin) based on whether it sees the market going up or down. Last week, I realized that a number of the more prevalent coins follow a similar path, wherein the percentages up and down from the previous price on a minute to minute basis match each other rather nicely. I have determined that from that, one could follow a number of these coins and watch for when one were to over-adjust and increase or decrease in price more than the rest. That would be an opportune time to buy/sell it for a higher/lower priced crypto in the same group, as it is overwhelmingly probable that the momentary divergence will resolve itself via the odd coin out rejoining the rest. Today, I have put that theory to test, albeit with some interesting results. Testing on the data I have collected over the past week, I found that holding a predetermined amount of crypto will cause a net loss of about 24% (it was a bad week for the market). I have found that using this strategy, my trader would only incur a net loss of about 12%, while keeping all funds in cryptocurrencies the entire time. I am collecting further data to hopefully test this theory again soon on a data set that either maintains its value over the duration, or slightly increases overall to see if the trend holds.