While building the project, we faced several technical and strategic hurdles:

⸻

🛠 Mainnet-only environment
Socios’ Fan Tokens are only available on the Chiliz mainnet — no testnet or developer sandbox is provided.
This forced us to deploy and test our smart contract directly on mainnet, making debugging slower and riskier.

⸻

🔒 No token fragmentation
Fan Tokens on Socios are non-fractional by default (i.e., you can’t send 0.5 token).
This is incompatible with micro-predictions or low-stakes interactions.
To solve this, we implemented wrapped fractional Fan Tokens, backed 1:1 by the real ones, to enable smaller predictions while staying economically sound.

⸻

⏱ Timing sync across users
Because our predictions are live and short-term (e.g. “Who will score next?”), it was crucial that all users watching the same stream receive the exact same challenge at the exact same moment.
To achieve this, we used WebSocket communication between the browser extensions and our backend, ensuring real-time broadcast of each challenge across all clients. That's why we have two repo.

⸻

⚠️ Transaction timing conflicts
Another major issue: blockchain transactions are not instant. If a user clicks “Predict” but doesn’t sign the transaction before the prediction timer ends (e.g. 5 seconds), it can break the fairness of the system.
We solved this by:
• Enforcing a prediction cutoff window (ex: predictions only accepted within 4 seconds)
• Validating on-chain timestamps, discarding late transactions at the smart contract level

⸻

🧠 Real-time match interpretation

We built an AI layer to parse live match logs and automatically:
• Suggest relevant predictions based on the in-game context (score, round, killfeed, etc.)
• Detect and validate outcomes fairly, without manual scoring

However, in the current version, AI suggestions and prediction results still require human validation.
This ensures that predictions match the actual game being streamed, and that outcomes are confirmed accurately.
We’re working towards full automation, but context alignment between stream and data source still needs human oversight.

⸻

🔐 No official access to game data

We don’t yet have official licensing agreements with game publishers (like Riot Games or Psyonix), so we rely on public APIs or overlays.

For example, we use Riot Games’ API to retrieve match logs — but face a 1–2 minute delay between the live action and the data becoming available.
This delay significantly impacts our ability to trigger or close predictions in real time, and limits the precision of automatic validations.

Overcoming these latency issues — while staying within the limits of public infrastructure — remains one of the biggest technical challenges.