MANTIS (SN123) Rolls Out New Trading Framework Focused on Market Dynamics

MANTIS (SN123) is replacing TRADE-MIX with a mechanism grounded in market microstructure theory, moving miners away from price prediction and toward modeling how markets are made.

The subnet generated over $60,000 in returns from an initial $28,000 account balance between January and late May, then took drawdowns when the market dynamic changed. That experience told the team point estimation was the wrong problem to solve.

Alongside the mechanism update, buybacks pause temporarily, a new communications manager joins, and a technical paper on the redesign is coming soon.

Why Point Estimation Failed

Hundreds of experiments in supervised learning confirmed the same conclusion: predicting price from historical data is the wrong move.

1. Noise is the market: Fischer Black showed in 1986 that markets require noise traders to exist. Liquidity depends on participants who are wrong.

2. Signal and noise are inseparable: Rational traders can never be certain whether others are informed or trading on noise.

3. Prices react asymmetrically to news: Bad news raises uncertainty and lowers expectations, and good news raises expectations but gets offset by increased uncertainty. Simple signal-to-price models break on that nonlinearity.

4. Most economic variables are unobservable: Supervised learning approaches keep failing because they extract signal from a process that is mostly noise by construction.

Liquidity providers guide price toward liquidity pools to unwind inventory, making market microstructure a function of structural dynamics rather than pattern recognition.

How the New Mechanism Works

The redesign replaces point estimation with fully specified trades that reward structural understanding.

1. Miners produce complete trades: Direction, Kelly-staked confidence, stop loss, and two take profit targets. Resolved against realized price paths across multiple timeframes.

2. Validators identify signal sources: Determining which sources are informed by microstructure and which are built on leveraged beta.

3. Skin in the game during regime change: Miners get differentiated by how their confidence quantification holds up when markets shift.

4. Signal decay gets priced in: Miners are incentivized to quantify exactly when their signal is decaying.

5. The $SN123 becomes essential to mining: The full detail lands in the upcoming technical paper.

The design came from months of conversation with miners, traders, and analysts pushing the team toward microstructure as the foundation.

Team, Communications, and Buybacks

Three operational changes land alongside the mechanism.

1. New communications manager hired: Weekly community updates and consistent transparency going forward.

2. Team expansion likely: An additional engineer or researcher may join as research produces more sophisticated approaches.

3. Buybacks paused: Every resource goes into the mechanism upgrade. Buybacks resume once the flagship trading strategy is upgraded, confirmed, and monetized.

How Markets Are Made, Not Where Prices Are Going

Mantis is pushing toward a trading product that gives SN123 holders direct utility through access to live signals, with access gated by token holdings.

The mechanism upgrade positions the subnet to extract signal from an angle no decentralized project has attempted, treating the question as how markets are made rather than where prices are going.

The next month or two will show how the software works, how it involves the best miners, and how it rewards them for adding predictive value.