A new arbitration possibility is the so-called input and output is performed in the same exchange. More specifically, although it is bought and sold simultaneously on two exchanges as in the scenario above, this operation is always followed by a complimentary one, of sale and purchase at the same time. Assuming that the movement of the pair on the two aforementioned exchanges is not perfectly synchronous, there are times when that difference is small and times when it is large. With the immediateedge bot this is the best deal that you can now have.
The Right Bet
The bet, in this case, is that the difference movement will be cyclical. A very simple algorithm is, for example, buying and selling when the difference is below the 5th percentile of the distribution of possible values and selling or buying when the difference is above the 95th percentile. In this case, it must be analyzed whether this difference exceeds trading fees and related risks, but it is no longer necessary to run the money between exchanges actions that come with additional costs.
The Different Types
These types of arbitrage are an example of simple market-neutral strategies, in which profits and losses are not affected by the global price movement relative to the base currency, ETH in example, but only by the evolution of the price difference. As expected, since it is quite easy to build such algorithms and the risks are relatively small, the market is quite crowded, and the more algorithms that do this, the lower the profit opportunities. Speculative arbitrage algorithms have a positive role, which is to equalize prices between exchange plants and increase liquidity, facilitating the meeting between supply and demand even if traded in different places.
The Other Catagory
Another fairly large category of algorithms that can be built is those that track the differences between currency pairs. In the case of these algorithms, the differences between two instruments are tracked and the bet is that the historical relationship between the price evolutions of the two will continue in the future.
- As with the second mode of arbitrage, the difference between prices or cumulative returns is the one relevant to making the decision to buy and sell two cryptocurrencies in different quantities simultaneously, so that the two pairs are hedged., that is, losses on one currency are offset by gains on the other, if the whole market moves in a certain direction.
- There are both classical methods, such as co-integration to determine the nature of the price relationship and the use of linear regression or filters on historical data to determine the hedging factor. Moreover, Machine Learning algorithms can be used to detect outliers and times when the relationship between two tools is at an extreme from which it will return, without breaking the existing historical link.
Example of the evolution of a pair of cryptocurrencies at the level of days
It is important to remember that it is difficult to find an algorithm that works, but at the same time the search space is infinite and things that seemingly have no meaning could prove useful. The map is not the territory, any metric that reduces the complex activity of the exchange-simplifies not allowed much, but at the same time any metric can be a projection of information and something simple may prove profitable in many cases so as to cover the losses when the decision is wrong.