Heuristics

Helio's Heuristic Approach

Core Heuristics

Helio bases itself on several heuristics, some are proprietary to the algorithm, but consist of looking for mints associated to multiple wallets existing in the same network. Mints with no relation to multiple nodes are removed in order to avoid node clustering with unnecessary data.

Algorithm Components

Cost Function

The algorithm being built employs a cost function, based on the market caps of mints in the network. It attempts to predict the value of a "pumped" coin, based on the patterns the wallets in the network employed in previous schemes.

Search Implementation

Because our search algorithm has no limit on depth (but is instead limited on total nodes), it is able to filter "unpromising" paths and focus on paths that are considered high-value. This allows the algorithm to find patterns even in CEX data.

Pattern Recognition

Because CEX's and DEX's have fairly low volume over small periods of time, entry and exit transactions can be matched and patterns can easily be discovered.

Future Improvements

Of course, it is theoretically impossible to achieve perfect information, but over time, the amount of data in the platform will increase, and thus our algorithm will have more data to be trained on.

Training Process

Our algorithm is also being consistently trained on "interesting" wallets that are submitted by our team.