Track Record

What this page answers

If GLCI says "loose" or "tight" right now, does history show that certain assets usually move a certain way over the next few months? And is that pattern strong enough to actually bet on? That is the whole point of this page.

1. The three cards below

• Strongest positive edge: the one asset + regime + horizon combo where the GLCI signal helped the most historically. "+5pp" means: when GLCI was in that regime, buying that asset for that many weeks had a hit rate 5 percentage points higher than the base rate (just buying on any random week). Bigger edge = stronger signal.
• Largest negative edge: the mirror image. Being in that regime hurt returns. Useful as a "when GLCI says X, don't touch Y" marker.
• Current GLCI regime: what the signal is flashing today, with the NFCI baseline alongside for comparison.

2. The Regime History bar

Each thin vertical stripe is one historical week, colored by the regime GLCI was in that week.

• Green = loose (liquidity easing, historically risk-on friendly)
• Yellow = neutral (nothing extreme either way)
• Red = tight (liquidity contracting, historically risk-off)

The three big boxes under the bar show what share of history sat in each regime.

3. The Forward Returns table (the main tool)

One row per asset (S&P 500, Gold, Bitcoin, etc.). Each row has 9 cells grouped by regime (Tight / Neutral / Loose) and horizon (4, 13, 26 weeks forward).

Inside each cell:

• Big number: median return over that horizon, starting from a week that was in that regime. Green = positive, red = negative.
• Small line: hit rate (fraction of windows that ended positive) and the edge in percentage points versus that asset's unconditional base rate. Positive edge = the regime label genuinely helped.
• Hover a cell: reveals the IQR (middle 50% of returns) and the 95% bootstrap confidence intervals on the median and hit rate.
• Empty cell with just "n=X": fewer than 20 observations in that bucket. Too thin to trust, so nothing is shown.

4. How to actually use this for positioning

• Scan the Loose columns for strong green numbers. Those are the assets and horizons that historically tended to work when liquidity was easing.
• Scan the Tight columns for deep red. Those are your "when GLCI is tight, this asset has suffered" warnings.
• Switch to the NFCI tab and compare. NFCI is a simpler, well-known stress index. If GLCI's edges don't beat NFCI's, the composite isn't adding information, and you could just use NFCI alone.
• This is a backtest. Past regimes may not repeat. The point is to calibrate your expectations, not to guarantee outcomes.