Portfolio Performance: First Half of 2026

What a full cross-asset Pythia portfolio did over the first half of 2026.

Last month we looked at what happens when you combine a handful of signals: equity index blends, metals, and a mix of the two. The takeaway was that combining signals reliably improved at least one dimension of risk-adjusted performance, and that cross-asset combinations did the most.

This month we run the same exercise on June's numbers, now adding the two new FX signals (the Euro and the Japanese Yen) for the full eight-signal set.

The short version: an equal-weight portfolio of all eight signals produced a Sharpe of 2.13 and a return of +4.27% over the first half, with a maximum drawdown of just -2.34%, matching the Sharpe of the single best signal at a fraction of its volatility.

The full portfolio: all eight signals, equal weight

Composition: S&P 500, Nasdaq-100, Russell 2000, Dow Jones, Gold, Silver, Euro, Japanese Yen, at 12.5% each.

Table 1: Eight-signal portfolio performance, first half 2026

Table 2: Eight-signal portfolio monthly returns (%)

Two things stand out. First, the portfolio's Sharpe of 2.13 sits right next to the best individual signal in the set (Silver at 2.14), but Silver carries 30.4% volatility and a -6.5% drawdown to get there, while the portfolio does it at 4.2% volatility and a -2.3% drawdown. Second, that -2.34% max drawdown is smaller than every signal in the book except the single lowest-volatility one (Japanese Yen at -1.96%), and less than half the drawdown of the large-cap equity or metals signals.

The portfolio is not the highest-returning configuration; that isn't what breadth buys you. What it buys is a return stream that stays positive through a first half in which four of the eight signals were individually negative.

Portfolios by asset class

Equity index (S&P 500, Nasdaq-100, Russell 2000, Dow Jones, 25% each)

Table 3: Equity index portfolio performance

Metals (Gold 50%, Silver 50%)

Table 4: Metals portfolio performance

FX (Japanese Yen 50%, Euro 50%)

Table 5: FX portfolio performance

Four observations

Breadth compresses risk. Eight signals equal-weighted produced the lowest volatility (4.19%) and one of the tightest drawdowns (-2.34%) of any return-generating configuration we tested, while matching the Sharpe of the single best signal at roughly one-seventh of its volatility. Adding structurally different signals doesn't just average returns; it shrinks the variance of the combined stream.

Cross-asset breadth still does the heavy lifting. Equity index signals were net negative in the first half, yet the blended portfolio stayed clearly positive because the metals and FX signals aren't driven by the same order-flow regime. The step from a single asset class to three is worth more than any reweighting inside one of them.

FX earns its seat as a diversifier, not a return driver, in this window. Standalone FX was roughly flat YTD. But its 1.97% volatility and -2.01% drawdown are the lowest in the set, its trailing-twelve-month Sharpe is 1.10, and adding it (with Dow Jones) to the blend lowers portfolio volatility and drawdown even though it contributes little return. That is what a diversifier is supposed to do.

Weighting matters, but direction matters more. These portfolios use simple fixed weights chosen for volatility balance, not optimized in-sample. The eight-signal book is a flat 12.5% each. The point isn't the exact numbers; it's that even an unsophisticated equal-weight combination produces a measurable risk-adjusted improvement when the signals carry low inter-asset correlation.

One caveat worth stating plainly: this is six months of data, and it's uneven. Metals returns are concentrated in the first quarter, equity signals had a weak first half, and the FX signals have the shortest live history of the set. These figures are directional, not stable estimates. The structural point holds regardless of the single-window magnitudes: because Pythia produces independent signals across equities, metals, and FX from a single model and API, this kind of cross-asset portfolio construction is possible at all, and it's the low correlation between those signals that improves the risk-adjusted profile. The specific magnitudes should be read with the caveats above in mind.

Where to go next

Individual signal benchmarks for S&P 500, Nasdaq-100, Russell 2000, Dow Jones, Gold, Silver, Euro, and Japanese Yen are published at quantumsignals.ai/benchmarks. Each report includes monthly Sharpe, return, win rate, max drawdown, and Calmar.

All results are out-of-sample backtests using pre-trained signals under realistic transaction cost assumptions. Past performance does not guarantee future results. Full methodology is described in our benchmark trading strategy blog post.