Instrument-Tier Calibration
PulseGrid v2.0 introduces instrument-tier calibration, recognizing that different asset classes absorb information at fundamentally different speeds. A geopolitical event affecting AAPL (mega-cap, highly liquid, widely covered) will be priced in within hours, while the same event affecting a commodity ETF or cryptocurrency may take days or weeks to fully reflect.
Theoretical Foundation
This calibration is grounded in the Adaptive Markets Hypothesis (Lo, 2004), which posits that market efficiency varies across instruments and time periods based on the population of active participants and their information processing capabilities. Empirical studies by Ding et al. (2015) confirmed that event-driven signals have different half-lives across asset classes.
Tier Definitions
PulseGrid classifies instruments into seven tiers:
| Tier | Examples | Decay Half-Life | Optimal Horizon | Sensitivity |
|---|---|---|---|---|
| Mega-Cap | AAPL, MSFT, GOOGL, AMZN | 5 days | 5 days | 1.0x |
| Mid-Cap | JPM, XOM | 7 days | 10 days | 1.1x |
| Sector ETF | XLF, XLE, XLK, XLV | 10 days | 10 days | 0.9x |
| Index | ^GSPC, ^DJI, ^IXIC | 7 days | 5 days | 0.8x |
| Commodity | GC=F, CL=F, SI=F | 10 days | 20 days | 1.2x |
| Forex | EURUSD=X, GBPUSD=X | 14 days | 20 days | 0.7x |
| Crypto | BTC-USD, ETH-USD | 3 days | 5 days | 1.5x |
Calibration Parameters
#### Decay Half-Life
The time-decay function uses an exponential decay with tier-specific half-life:
> Decay(t) = exp(-ln(2) * t / half_life_tier)
Where t is the number of days since the event occurred. Mega-cap stocks use a 5-day half-life (events lose half their influence in 5 days), while commodities use 10 days (events persist longer due to slower information absorption).
#### Optimal Prediction Horizon
The multi-horizon validation system (P1) uses tier-specific optimal horizons. Instead of validating all instruments at the same 1-day or 5-day horizon, each tier is validated at its empirically optimal window:
- Mega-cap (5d): Highly liquid, events priced in quickly
- Mid-cap (10d): Moderate liquidity, slower price discovery
- Commodity (20d): Supply/demand imbalances take weeks to manifest
- Crypto (5d): High volatility but rapid information absorption
#### Sensitivity Multiplier
The sensitivity multiplier scales the raw PG score to account for inherent volatility differences:
> Adjusted Score = Raw PG Score * sensitivity_tier
Crypto instruments (1.5x) have amplified scores because their higher baseline volatility means the same event magnitude produces larger price moves. Forex instruments (0.7x) are dampened because currency pairs are inherently less volatile.
Tier Classification Logic
Instruments are classified by a combination of:
1. Symbol pattern matching: Forex pairs (contains "="), crypto (contains "-USD"), futures (contains "=F")
2. Prefix matching: Index symbols starting with "^" or "."
3. Explicit mapping: Known mega-caps, sector ETFs, and commodities are explicitly listed
4. Default: Unrecognized symbols default to mid-cap tier
Frontend Display
The instrument tier is displayed as a badge on the instrument detail page header, next to the instrument name. The tier classification helps users understand why different instruments may have different score magnitudes and validation horizons.
Limitations
- Tier classification is rule-based, not dynamically computed from market microstructure data
- The sensitivity multipliers are calibrated from historical averages and may not reflect current market conditions
- Some instruments may be misclassified if their characteristics have changed (e.g., a mid-cap that has grown to mega-cap)
- Cross-listed instruments may behave differently in different markets