Devoured - April 21, 2026
How Freeport Users Made 11.7% on $27M in 45 Days (6 minute read)

How Freeport Users Made 11.7% on $27M in 45 Days (6 minute read)

Crypto aifintechtrading Read original

A new AI-powered trading platform reports 11.7% user returns in 45 days, but transparently attributes most gains to market beta and momentum rather than platform-generated alpha.

What: Freeport is an AI news feed platform with one-tap trading execution that generated $27M in volume across its first 45 days. Users primarily traded major macro events: oil futures during Middle East escalation (WTI crude from $60s to $100+) and NASDAQ longs during subsequent de-escalation, achieving 11.7% aggregate money-weighted returns.
Why it matters: This represents unusual transparency for a fintech product launch. Rather than claiming credit for returns, Freeport breaks down attribution: 4-5% from market beta (being long during a rally), 3-4% from momentum (following established trends), 2-3% from oil concentration, leaving only 2-3% residual that isn't statistically significant at 46 days. The counternarrative to typical fintech hype makes the underlying mechanism more interesting—their data shows top performers (18.2% returns) traded less frequently and used lower leverage than average users.
Takeaway: If building trading or investment tools, consider optimizing for user outcomes over engagement—Freeport's best traders averaged 2.1 trades per day versus 5.8 for median users, held positions 31 hours versus 19, and used 2.4x leverage versus 3.3x.
Deep dive
  • Platform launched February 28, 2026 during U.S.-Israel strikes on Iran that killed Supreme Leader Khamenei, creating immediate opportunity to test event-driven trading thesis
  • Users traded primarily real assets (NASDAQ/S&P indices, WTI crude, crypto) at 2-4x leverage, far below 10-200x platform maximums, suggesting measured risk management rather than speculation
  • Most oil profits came not from initial headline trades but from users entering hours/days later as escalation deepened, potentially exploiting serial correlation in geopolitical events similar to post-earnings announcement drift
  • When diplomatic channels reopened, users went long equities rather than shorting oil, possibly due to feed architecture biased against shorts in favor of anti-correlated longs, aligning with research showing negative media content overstates actual deterioration
  • Return decomposition via factor regression: approximately 4-5% from market beta (users were long during rally), 3-4% from momentum (following recent price moves), 2-3% from concentrated WTI exposure, 2-3% unexplained residual
  • The 2-3% residual has t-statistic of ~0.3 (not statistically significant) but would be economically meaningful if sustained, as even 1% over 45 days would annualize above most hedge fund performance
  • Top 1% of users (25 traders) achieved 18.2% returns through counter-intuitive behavior: traded less (2.1 vs 5.8 daily trades), used lower leverage (2.4x vs 3.3x), held longer (31 vs 19 hours median)
  • Platform cites academic research to support mechanisms: Hong and Stein (1999) on event-driven momentum from slow information diffusion, Tetlock (2007) on media negativity bias, Jame et al. (2022) on curated analysis improving retail order flow predictiveness
  • Product philosophy explicitly rejects engagement optimization in favor of outcome optimization, implementing fewer notifications and context for non-trading as much as trading signals
  • Volume breakdown shows institutional-style positioning: 33% in equity index perpetuals, 15% WTI crude, with remaining volume across single stocks, crypto, and pre-IPO tokens
Decoder
  • Money-weighted returns: Returns calculated by weighting each position by its dollar size, giving more influence to larger trades (versus time-weighted returns that treat all periods equally)
  • Market beta: The portion of returns explained by broad market movement—if the market rises 5% and you're long, you capture that regardless of skill
  • Momentum: Persistent tendency for assets that have risen recently to continue rising, one of the most robust empirical patterns in finance across decades of data
  • Factor regression: Statistical technique decomposing returns into systematic components (beta, momentum, etc.) versus unexplained residual that might represent skill or luck
  • T-statistic: Measure of statistical significance; values below ~2.0 suggest results could easily occur by chance, Freeport's 0.3 indicates their residual returns are not statistically meaningful yet
  • Post-earnings announcement drift (PEAD): Phenomenon where stock prices continue drifting in the direction of an earnings surprise for weeks afterward due to underreaction
  • WTI crude: West Texas Intermediate, the U.S. benchmark for oil pricing
  • Perpetual futures: Crypto-style derivative contracts with no expiration date, maintained through funding rate mechanisms
Original article

Freeport, an AI news feed platform with one-tap trading execution, reported $27M in volume and 11.7% aggregate money-weighted returns across its first 45 days, with users trading NASDAQ, S&P, crude oil, and crypto at 2-4x average leverage. The platform's two dominant macro trades were a WTI crude long from the low $60s to above $100 on Middle East tensions, followed by a NASDAQ 100 long that captured about 15% off March lows as diplomatic channels reopened. Return attribution assigns 4-5% to market beta, 3-4% to momentum, and 2-3% to oil concentration, with 2-3% residual alpha that lacks statistical significance at 46 days. Top users (1% of the base, 18.2% returns) averaged 2.1 trades per day versus a 5.8 median, held positions 31 hours versus 19, and used 2.4x leverage versus 3.3x.