US stock universe: Top 1200. Daily returns are volatility adjusted, PC1-3 are first 3 principal components from a 3yr daily covariance matrix. Consensus Wts are an average of "artificial intelligence" thematic ETFs, with non-consensus wts for each ETF being the difference in holdings vs consensus

Standard 5 tier signal, 20bps TCs, signal rebalanced weekly/biweekly/monthly. Cash portfolio int rate = 0

Standard 5 tier signal, 20bps TCs, signal rebalanced weekly/biweekly/monthly. Cash portfolio int rate = 0

Standard 5 tier signal, 20bps TCs, signal rebalanced weekly/biweekly/monthly. Cash portfolio int rate = 0

20bps transaction costs, simplified signal: up trends 1, down trends 0, neutral 0.5; signal rebalanced on Wednesdays

20bps transaction costs, simplified signal: up trends 1, down trends 0, neutral 0.5; signal rebalanced on Wednesdays

2Step position constraint approaches allow for 2 different weights for Tier 1, Tier 2 position concentrations

Portfolio Weights over time

Conditioning Mean Reversion strategies on the Bitcoin Trend Indicator. If BTI values are <0.5, allow short term reversal strategies to trade. Otherwise 0 out in the event BTI is 0.5 or 1 values

Portfolio optimized constraints: +/- 0.3 beta to S&P500, 10yr Bond, gross leverage 4x, position limits by asset based on long term volatility.

Optimized portfolio of forecasts vs next period return and position level attribution. Portfolio optimized constraints: +/- 0.3 beta to S&P500, 10yr Bond, gross leverage 4x, position limits by asset based on long term volatility. T1 is 1day implementation trading lag, T2 is 2day lag.

5bps TCs, 2 period signal lag, 100% gross weights

5bps TCs, 2 period signal lag, 100% gross weights

5bps TCs, 2 period signal lag, 100% gross weights

5bps TCs, 2 period signal lag, 100% gross weights

5bps TCs, 2 period signal lag, 100% gross weights

5bps TCs, 2 period signal lag, 100% gross weights

Long/short overlay based on correlation rank of ARP strategies against 60/40 benchmark portfolio

realized_vol_window <- 60 PCA_Cov_EMWA_window <- 24*30 1hr returns EWMA Coviance

Internal R code vs Springworks Python

Comparison of long only Sector ETF model (implied by Shalom Global weights provided) versus internally generated strategy utilizing Momentum / Reversal / Value / Low Beta signals versus SP500 ETF

Comparison of long only Sector ETF model implied by Shalom Global weights sent versus internally generated strategy utilizing Momentum / Reversal / Value / Low Beta signals

Return contributions by underlying ETF proxies. Sector model has dominated performance over period, while Bond (AGG) and international equity (EFA) allocations have underperformed

% of Capital Weights over time, Vor1 103121 model

Comparison of 08-04-21 Algo with 10-13-21 revision

Strategy signal blend across daily, 6hr, 1hr, 15min signals across BTC, ETH, ADA, LTC, BCH, DASH, ZEC tokens. Transaction costs tested at 0 - 2bps

Active strategy does not short crypto assets, minimum weight is 0%, maximum weight is 2x index weight Jan 2021 Bloomberg/Galaxy weights: Bitcoin: 20% Ethereum: 20% Litecoin: 9% Bitcoin Cash: 8% EOS: 3%

Hypothetical performance of Bloomberg/Galaxy Crypto Index with active quant strategy overlay. Annual Returns BBG/Galaxy: 126% Active Fund Hypothetical: 177% Alpha TE: 51% Risk Adjusted Performance (Return per Volatility) BBG/Galaxy: 1.46 Active Fund Hypothetical: 1.91 Alpha TE: 1.53

Bitcoin has shown empirically to be a source of diversified returns, with weekly correlations within (+/-) 0.4 across liquid multi-asset benchmarks, since 2015. There's been a notable increase in correlation between Bitcoin and the S&P 500, Gold and the US Dollar (decrease) after the March 2020 market drop. This could potentially be explained by the increase of institutional investment into Bitcoin, in response to the increased fiscal and monetary support for the economy during the COVID-19 epidemic. (Source: BTCUSD data from CoinMarketCap; Broad asset class proxies from Yahoo Finance; Weekly return correlations calculated using EWMA smoother, 26w half-life)

Ln($/sqft)

Regions include CA, OR, WA, NV, AZ, CO and HI

From 200k home price sales (primarily from CA). School index is from 2011 year, weather data from 1990

found a bug in the previous published chart

Slow signal (10d horizon), 10/23/2020

Current signal (1d horizon) for 10/23/2020

Current positioning of directional signals