Standard macro backdrop plus EvolveCapital's own regime signals and the full asset universe by category.
Standard macro backdrop from FRED (plus WTI crude & gold from market prices), refreshed nightly. Change is versus the prior reading, in the series' own units (pp = percentage points). The 1M/3M/6M/12M chips below each chart show that horizon's change, its historical percentile rank, and (hover for detail) the S&P 500's forward 12-month return historically when that indicator was in the same percentile bucket — shown only where there's a large enough historical sample.
Central bank balance sheets (Fed, ECB, BOJ) plus a "Private Liquidity" group, with the same current-value/1M/3M/6M/12M-change/percentile/forward-S&P-500 treatment as the Macro tab, plus a t-test of each bucket's forward-return mean against the unconditional (all-history) mean — a * marks a statistically significant difference (p<0.05). Bank of England and the PBoC aren't included: BOE's only current-through-today FRED series is discontinued or historical-only (1701–2016), and the PBoC has no clean free current series on FRED at all — rather than substitute unreliable or scraped data, this covers 3 of the 5 "major" central banks. "Private Liquidity" approximates (does not replicate) Michael Howell/CrossBorderCapital's Global Liquidity Index private-sector component — his actual index is a proprietary weighted composite across ~80 countries; these are the closest free FRED proxies to his stated bank/shadow-bank-credit and repo/collateral-market inputs. "Equity Supply" is the Fed's (Z.1) net equity issuance by US nonfinancial companies — IPOs and secondaries minus buybacks and cash-M&A share retirements, quarterly at an annualized rate; negative means more stock was retired than issued. It doesn't include lock-up unlocks (no free aggregate source) or financial companies.
VIX, MOVE and GVZ are real implied-volatility indices (equity, Treasury, and gold respectively), free via Yahoo Finance. There's no free equivalent to Deutsche Bank's CVIX (currency vol) or a credit-implied-vol index, so those rows — and the TLT row, since MOVE has no free realized-vol counterpart otherwise — are realized-volatility proxies built from free price/spread data instead, marked "proxy" rather than passed off as the real benchmark. The "5Y level" chip is where today's raw reading ranks within that series' own trailing 5-year range (not a change, unlike the 1M/3M/6M/12M chips below it).
Where the economy sits in the cycle and what has usually come next: an inflation-vs-growth clock back to 1948, the order in which higher real rates have tended to hit markets, and S&P 500 seasonality since 1928. Refreshed nightly.
Clock: inflation is CPI (CPIAUCSL) and growth is industrial production (INDPRO), both year-over-year, from FRED; each axis is split into low / middle / high thirds of its own history since 1948 (cut-offs use the whole history, so they shift slightly as months arrive). Stock returns are the S&P 500 price index (^GSPC, no dividends), each month credited to the prior month's reading since CPI comes out mid-way through the next. Rates: Fed funds (FEDFUNDS, from 1954), the curve as 10-year minus 3-month Treasury yield (GS10 − TB3MS), the real 10-year as GS10 minus CPI inflation. Transmission: the trigger is the 10-year TIPS yield (DFII10); private credit is proxied by BDCs (BIZD); CCC and BB spreads are ICE BofA option-adjusted spreads (FRED, about 3 years of free history); utilities XLU, REITs VNQ; demand is real retail sales (RRSFS) and real PCE (PCEC96); earnings are corporate profits after tax (CP, quarterly). Seasonality: ^GSPC daily closes, forward 63 trading days. Averages describe history; they are not forecasts.
The exact regime signals EvolveCapital's own strategy filters already read — not generic indicators. See strategies/src/optimizations/dynamic.py and strategies/src/filter/top_performer_filter.py.
Inflation_Band is EvolveCapital's internal 0–5 classification of month-over-month CPI change (Supabase market_data.inflation_data). The only documented threshold is band ≥ 5 = "high inflation", which restricts strategies like theme_001 to Energy / International Equities / Infrastructure & Transportation / US Alt / Fixed Income that month (see filter_symbols_by_inflation_regime()). Bands 0–4 are internal percentile cutoffs not exposed by this API, so they're shown as a raw scale, not interpreted.
SMA bounce and deep-drawdown events for SPY (Supabase market_data.market_sma_events) — the same table _apply_market_regime_filters() reads to decide bounce-month restrictions (e.g. the Momentum/Growth/International-only rule for theme_001/factor_001).
| Event | Date |
|---|
Every ticker in Supabase's market_data.asset_category_mappings, grouped by category (a ticker can appear in more than one), with trailing compounded return by period. Not scoped to any single strategy's universe.
| Ticker | 1M | 3M | 6M | 12M |
|---|
The 11 GICS S&P 500 sectors (Select Sector SPDR ETFs), trailing return by period. Distinct from the By Asset Class tab's Supabase-driven category taxonomy — this is the standard 11-sector breakdown. Sectors are ranked by risk-adjusted momentum: the average of the 3M, 6M and 12M returns divided by annualized volatility over the last 12 months, so a steady climb outranks an equally large but choppier one.
| # | Sector | Ticker | Risk-adj. momentum | Vol (12M) | 1W | 1M | 3M | 6M | 12M |
|---|
Every US-listed ETF (Nasdaq's daily symbol directory, all exchanges) ranked by trailing return, refreshed nightly. Returns are over trailing trading days (1D = 1, 1W = 5, 1M = 21, 3M = 63, 6M = 126, 12M = 252), price plus reinvested distributions. AUM is the fund's net assets as Yahoo reports them; Avg $ vol is the average dollar value traded per day over the last 20 trading days, and funds under $1M a day are left out. Leveraged and inverse funds, and single-stock funds (leveraged or option-income funds on one company, e.g. YieldMax MRNA Option Income), are flagged from their names and hidden unless you include them. Unrecorded splits and one-day bad prints are corrected before ranking.
Median return by calendar month, trailing 10 years only (median is more robust to a single outlier year than average; 10 years is more regime-relevant than three-plus decades back). Month-of-year rather than week-of-year: same sample depth per bucket either way, but week-of-year adds calendar-alignment noise from 52/53-week years. Click any month header below to sort by it; the current month is highlighted. Hover a cell or bar for the full breakdown (median, average, win rate, years of data).