No-code trading bot: how it decides how much to buy

A no-code trading bot decides how much to buy with the size of its forecast, not the direction the forecast points. In Finforge's Sophon-3 paper, Version 1.0, published in July 2026, a step called Make Portfolio turns each forecast into a position, and a minus 5% forecast and a plus 5% forecast both earn a 5% weight. The magnitude sets the bet. The sign is thrown away.

That is the most counterintuitive decision in the paper, so it helps to see the machine it runs inside. The no-code chain screens a universe, a Forecast node estimates what each name does next, and Make Portfolio writes the orders. The bot is long-only and unlevered, meaning it never borrows money to buy more, and there is no portfolio optimizer on top. It is always fully invested, with cash weight at or below 0.1%.

How a no-code trading bot turns a forecast into a stake

The sizing rule is one line of arithmetic. The raw size of a position is the absolute value of the forecast, and each weight is that raw size divided by the total across the whole set. Absolute value is what strips off the minus sign. So a -5% guess and a +5% guess put the same 5% into the position, and only the strength of the guess controls the stake.

If every magnitude collapses, the bot gives up on sizing and falls back to equal weight across the names it still holds. There is a default rather than a crash.

Why a bot bets on size, not on direction

Ignoring the call while trusting its strength is a claim, and the paper says the numbers back it. When the same agent was forced to equal-weight its positions, the forecast-magnitude sizing it replaced added 23% to 43% of active return a year across the three agents, pooled at +36% a year with a t-value of 3.8, significant in all three. A t-value is a rough signal-to-noise score: the higher it is, the less likely the gap is random luck.

When keeping only the up-calls backfires

A no-code bot could keep only the forecasts that point up and bet on those. The paper tried it, and the result is honestly mixed. Positive-only sizing made two of the three flows worse, once significantly by 18% a year, and improved one. The absolute-magnitude default stays, not because it always wins, but because it lost least on the net evidence.

The paper also says flat out that it does not yet know why sizing works. One working idea: a model trained on returns sends big forecasts to names with wide swings, whichever way the sign points, so magnitude may be proxying volatility among the stocks the screener kept rather than voting for direction. The August 2026 revision plans two diagnostics to settle it, whether magnitude tracks realized volatility and whether it beats a plain volatility proxy on the same names.

Getting the money in, the other half of a no-code bot

Sizing sets the stake. Execution has to place it, and here the paper targets weights, not prices. The objective is to make the realized weights match the intended weights and keep weight drift small, and it judges execution quality on exactly that. A decision made at the close of one session is placed about five minutes before the next close, on highly liquid U.S. large-caps at retail scale.

The paper does not claim slippage is zero, only that it is small at this size, because the daily rebalances are small and the names are among the most liquid U.S. large-caps. Alpaca, the base broker, has been commission-free on U.S. equities since 2018 and is used for both backtest and live. These figures are model comparisons from the paper, not an account statement, and past performance is not a guide to future returns.

Questions people ask

Why would a bot throw away the direction of its forecast?

Because the measured value sat in the size, not the call. Magnitude-based sizing added 23% to 43% of active return a year across the three agents, pooled at 36%, while keeping only the up-calls made two of the three flows worse.

Is a no-code trading bot a DIY strategy builder?

Not at Finforge. Finforge is not a no-code builder where you assemble and test your own bots. It ships pre-built live agents instead, Sophon Apex, Core, Pillar and Surge, trading in public on founder capital in Alpaca paper accounts since April 21, 2026. The no-code promise is that the quant machinery runs for you, not that you build it. No customer money is traded before launch.

Finforge runs its Sophon agents in public and publishes every trade, benchmark and drawdown. Check the live results, benchmarks and drawdowns for every Sophon agent at finforge.com/agents, read the full method in the Sophon-3 research summary, and join the waitlist.