Welcome to the first installment of our series exploring advanced network analysis techniques used by Fyde. Eighteen months ago, when we began integrating AI into our protocol, we kept our methods undercover. Now, we’re ready to deepen the community’s understanding of the intersection between crypto and AI through the lens of meme coin trading by sharing insights from our developed tools.

In this series, we will cover Network Analysis and its application to power Graph Machine Learning (GML) for analyzing tokens from a trading perspective.

What is Network Analysis?

Network Analysis is a methodical approach used to explore network structures—whether social, biological, or financial, such as blockchain. It involves mapping entities as ‘nodes’ and their interactions as ‘edges.’ This technique uncovers deep insights, identifies key influencers, and reveals behavioral patterns. For blockchain, it clarifies transaction flows and actor interplay, illuminating market trends, liquidity, and risks.

What is Graph Machine Learning?

Graph Machine Learning (GML) constructs models using data organized as graphs, with nodes as entities and edges as interactions. It focuses on prediction and decision-making by analyzing nodes, edges, and topology. For blockchain, GML enhances transaction analysis, fraud detection, and strategic optimizations, leveraging the complex dynamics of network interactions.

This analysis, alongside tools like the Fyde Simulation Engine and Rapid Arbitrage Calculation Engine – Rust System (RACE-rs), will remain proprietary but accessible to users within Fyde’s Liquid Vault.

Let’s start by applying these techniques to analyze two meme coins, Pepe and Pepe 2.0, which, despite sharing a name, offer vastly different risk/return profiles to investors.

We’ll explore why these coins were selected and how they reflect our investment strategies and cultural engagement with meme coins. This series will also cover other tokens, such as harrypotterobamasonic10inu, Bobo, and various non-meme coins wherever useful for illustrative purposes.

We invite you to share your favorite tokens for potential future analysis.

---

What is Pepe the Frog?

Pepe the Frog has become a widely recognized internet meme. Its popularity exploded in the 2010s, appearing across social media. In crypto, early iterations of Pepe appeared as NFTs, but the biggest success was Pepe token itself, leading to high volatility and stories like How Pepe Coin Made Me a Crypto Millionaire in Two Weeks.

This fueled a meme coin frenzy where speculation ran wild. Pepe 2.0 emerged among many derivatives, hoping to mirror the original’s success, though the outcomes and trading patterns differed widely.

Pepe:

• Token contract address: 0x6982508145454ce325ddbe47a25d4ec3d2311933
• Price Chart: CoinMarketCap - Pepe
• Launch Date: April 14th, 2023
• Nodes (Unique Wallets): 443,185
• Edges (Transactions): 2,832,228

Pepe 2.0:

• Token contract address: 0x0305f515fa978cf87226cf8A9776D25bcfb2Cc0B
• Price Chart: CoinMarketCap - Pepe 2.0
• Launch Date: January 2, 2024
• Nodes (Unique Wallets): 12,435
• Edges (Transactions): 50,145

---

How Does Fyde Analyze Blockchain Networks?

At Fyde, we leverage AI to enhance trading strategies and stress-test tokenomics. Our primary objective is to analyze wallets using network analysis to detect relationships and unveil the behavior patterns of key liquidity providers.

We use algorithms to identify outliers and evaluate risks, including centralization metrics and patterns indicative of wash trading. Our goal is to understand liquidity dynamics within token ecosystems deeply.

Some algorithms follow a time complexity of O(N^2 + E). Analyzing Pepe and Pepe 2.0 could require over 207 billion computations. Fyde optimizes performance by pruning unnecessary calculations and employing efficient data structures.

With GML, we construct detailed network graphs of wallets (nodes) and transactions (edges), enriched with metadata like address labels and transaction volumes. Our machine learning models use these graphs to predict behaviors, such as contract identification and transaction flows into protocols like Uniswap V3.

The proprietary Fyde Simulation Engine back-tests algorithms, simulating DeFi trading behaviors, optimizing arbitrage, and enhancing predictive models—all accelerated by parallel EVM technology.

Upcoming Analysis Series

In upcoming articles, we’ll showcase techniques for identifying significant wallets and analyze a notable MEV bot, offering insights into the profiles and behaviors of key tokens. Our approach focuses on “hidden gems” that traditional methods often miss.

Deep Dive into Liquidity

Liquidity is crucial at Fyde, and our Liquid Vault embodies this philosophy. We analyze liquidity pathways and decentralization to enhance risk management. Articles will provide insights on managing liquidity dynamics and decentralization, helping you make informed trading decisions.

Our tools include analyses of common liquidity pathways, decentralization metrics, and topology-based algorithms for volatility forecasting. This helps us assess liquidity control, crucial for understanding market manipulation risks.

---

We look forward to sharing deep insights into our network analysis methodologies. Follow along as we elevate the discourse and empower you with knowledge for informed trading decisions.

Follow Along:

Special thanks to Cryo and R-eth for enabling us to process blockchain data rapidly. For guidance on setting up a RETH node and using Cryo for blockchain analysis, check out these resources:

• Web3 and Rust Resources
• Setting up a RETH node
• Cryo Developer Documentation