RACS Portfolio Construction · Simulated Book

What the Backtest Portfolio
Looks Like

This is a simulated portfolio, not a live traded book. It shows how a strategy that mechanically follows the top RACS signals would have performed, after realistic transaction costs and execution constraints.

How this page is built
1
Score signals
RACS ranks every ticker-quarter by activist conviction × macro tailwind
2
Select top decile
Only the top 10% of scored signals pass into the simulated book
3
Size by risk budget
PortfolioConstructor weights positions by volatility target and sector caps
4
Backtest with costs
All trades simulated with T+1 fill delay, slippage, and ADV participation limits

How to Read This Page

Click any card for a plain-English explanation of the key metric.

◎
Sharpe Ratio
Risk-adjusted return
▾
Target: > 1.0
▽
Max Drawdown
Worst peak-to-trough loss
▾
Target: < −15%
⇄
Gross Exposure
Σ |position sizes|
▾
Target: ≤ 150%
α
Alpha (α)
Return beyond factor models
▾
Target: > 0%
⚠

Small sample. This run of the simulated book closed only 9 trades. Every figure below is directional, not statistically conclusive, until more filing history flows through the pipeline.

Annualised Sharpe
-0.594
Target ≥ 1.0
Total Trades
9
0 survivorship flags
Turnover
40%
Annualised
Open Positions
6
Filing lag 45d · fill delay 1d

Performance by Regime

Trade outcomes grouped by the HMM macro regime active at entry. p-values are raw, pre-FDR-correction.

Goldilocks
Trades3
Mean Return+8.6%
Sharpe0.88
Max DD-11.3%
p = 0.262not significant
Recovery
Trades6
Mean Return-19.8%
Sharpe-1.03
Max DD-66.5%
p = 0.869not significant

Top Holdings

Ranked by mean forward return across the trade ledger

#CUSIPMean Return
182846H405+27.49%
294419L101+18.36%
3361008105-5.88%
486384P109-11.26%
519260Q107-31.88%
681730H109-46.57%

Ticker did not resolve for these CUSIPs in this run; see the Provenance gate on the Validation page for the current CUSIP-to-ticker resolution rate.

Fama-French 5-Factor + Momentum Attribution

Regresses trade returns against systematic risk factors to isolate alpha.

⊘
Skipped: not enough data

8 of 9 ledger trades survived the factor-regression join. A regression needs materially more observations than regressors (6 factors) to produce a stable estimate, so this run does not compute alpha or R².

Capacity Analysis

Simulated AUM scaling: as capital grows, ADV participation limits exclude more positions from the book.

AUMPositionsExcludedSharpeMean Return
$10M633%-0.68-12.6%
$20M544%-0.62-12.9%
$40M367%not availablenot available
$79M278%not availablenot available
$158M278%not availablenot available
$316M278%not availablenot available
$631M0100%not availablenot available
$1259M0100%not availablenot available
$2512M0100%not availablenot available
$5012M0100%not availablenot available

Capacity cliff: by $631M AUM, participation limits exclude every current position from this simulated book.

Signal Decay

Information coefficient of the RACS score against forward returns, across holding horizons. Half-life: 5 days.

1d
0.103
p = 0.35 · n = 86
5d
-0.011
p = 0.92 · n = 86
20d
-0.017
p = 0.88 · n = 86
60d
0.036
p = 0.74 · n = 86