Retail Flow
"Retail Flow" cross-sectional factor is designed to measure and respond to retail investor activity.
By analyzing individual executed trades sourced from exchanges, the strategy identifies assets heavily influenced by retail participation. It then takes systematically contrarian positions, seeking to exploit predictable patterns of overreaction and herding behavior.
Order sizes and types are utilized to differentiate between retail and institutional activity. To quantify trade imbalances on the filtered trades, five distinct techniques are employed, their outputs are normalized and combined in an ensemble approach, without any optimization or parameter fitting.
The asset universe consists of the most liquid and actively traded assets, identified on rolling basis - various techniques employed to keep it both stable and relevant, as well as survivorship-bias free.
To balance each asset's risk contribution, positions are scaled according to the inverse of their rolling volatility.
The factor is available (point-in-time) with hourly updates, 5 minutes past the hour (UTC).
Download data
| Benchmark (BTC) | Portfolio | |
|---|---|---|
0.71 | 1.52 | |
27.6% | 55.4% | |
0.00 | 0.49 | |
1.00 | -0.01 | |
-76.6% | -36% | |
57.8% | 32.5% | |
0% | 36.2% | |
100% | 4% | |
100% | 201.7% |
Period: Nov 15, 2020 – Jul 20, 2026
Performance
Report Period Start Date Nov 2020 · End Date Jul 2026Cumulative Returns
MTD−2.5%
Last Month5.8%
1M−3.0%
3M26.7%
YTD31.7%
Annualised Returns (CAGR)
1Y21.7%
3Y37.9%
5Y38.1%
SI55.4%
Risk Profile
Realised Volatility (annualised)
1M15.7%
3M21.4%
1Y24.1%
3Y28.2%
5Y27.9%
SI32.5%
Max Drawdown
%−36.0%
Date2024-10-03
Monthly Returns Heatmap
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2020 | — | — | — | — | — | — | — | — | — | — | −4.8% | +36.1% | +29.6% |
| 2021 | +42.0% | −13.9% | +34.4% | +13.0% | +25.1% | −9.3% | −5.5% | +9.2% | +11.6% | +6.6% | +13.5% | −4.4% | +180.9% |
| 2022 | −9.2% | +0.5% | −0.1% | −11.1% | +6.5% | +21.6% | +2.8% | −3.7% | +13.7% | +5.8% | +6.8% | +10.0% | +46.8% |
| 2023 | +1.6% | +0.1% | −11.0% | −5.1% | −1.2% | +5.2% | −3.7% | +1.4% | +7.0% | +2.8% | +1.9% | +25.9% | +23.0% |
| 2024 | −7.9% | +10.4% | +32.0% | +11.2% | −13.9% | −3.6% | −2.4% | −6.3% | −10.9% | +10.9% | +12.2% | −10.0% | +13.3% |
| 2025 | +4.1% | +10.2% | +7.1% | +7.0% | +7.0% | +4.0% | −3.0% | +2.2% | −2.9% | −6.7% | +0.9% | −5.8% | +24.8% |
| 2026 | +7.6% | +10.4% | −10.8% | −2.7% | +23.7% | +5.8% | −2.5% | — | — | — | — | — | +31.7% |
Rolling Metrics & Exposure
Net Exposure of Portfolio
Last 30 days hidden — subscribe for live portfolio weights.
Sharpe Ratio
Rolling 12 MonthsBeta
Rolling 12 Months - Benchmark: BTCPortfolio Variants
Current Weights
7/20/2026, 10:02:41 AM (UTC)
The portfolio is designed to be market-neutral on a volatilty-adjusted basis, therefore it may have net directional exposure on any given day. The weighting mechanism ensures that volatility-weighted directional exposure is negligible on the portfolio level.
| Ticker | Name | Arrival Price | Weight | Date |
|---|---|---|---|---|
| BTC | Bitcoin | 64148.081 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| ETH | Ethereum | 1863.186 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| SOL | Solana | 76.412 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| XRP | Ripple | 1.09 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| DOGE | Dogecoin | 0.07200567849344407 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| SUI | Sui | 0.742 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| BNB | Binance Coin | 565.614 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| ADA | Cardano | 0.163 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| LTC | Litecoin | 46.698 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| LINK | Chainlink | 8.401 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| AVAX | Avalanche | 6.553 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| BCH | Bitcoin Cash | 212.683 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| HBAR | Hedera | 0.0659042156618648 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| TRX | TRON | 0.326 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| XLM | Stellar | 0.187 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| SHIB | SHIB | 0.000004201073419177 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| HYPE | HYPE | 60.704 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| XMR | Monero | 334.381 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| ZEC | ZEC | 530.564 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| GRAM | GRAM | 1.52 USD | 0.00% | 2026-07-20T10:02:41.195739+00:00 |
| 0.00% | ||||
Properties
Cross-Sectional | |
Derivative Exchanges | |
400+ Tickers | |
10 Day |
2024-01-01 | |
Daily | |
Hourly, 5 minutes past the hour (UTC) | |
Yes |
Frequently Asked Questions
What is the construction methodology of the model portfolio on display? How do I replicate it?
The portfolio is constructed by applying cross-sectional binning - quantiles. This ensures that the weights are calculated in proportion to the strength of the factor. The assets are then weighted according to the inverse of their rolling volatility, to mitigate the impact of widely different volatilities of digital assets. We provide a replication notebook in the "Resources" section that will produce a very similar portfolio than what's shown on the site.
Is the raw factor data available?
Yes, we serve the raw factor data via the "portfolio/factors" endpoint, please see the API Docs for more information.
What is smoothing? Some factors have by default setting (eg. 10 Day Moving Average) applied. Why is that?
Some of the cross-sectional factors have very high turnover, and despite being highly predictive, the alpha does not survive transaction costs. Smoothing (applying a simple moving average) is a way to reduce the turnover and the impact of transaction costs.
What is the universe of assets? How is it determined? Is it survivorship-bias free?
The universe of assets (there are multiple variants, for example, top 20, 30, 40 market cap digital assets) is the most liquid and actively traded assets, identified on rolling basis - various techniques (volume, open interest, volatility filters) are employed to keep it both stable and relevant. The universe is survivorship-bias free.
What is the rebalancing frequency?
The portfolio is rebalanced daily, with hourly updates, 5 minutes past the hour (UTC), and available point-in-time.
What are the transaction cost / slippage assumptions?
There are 0.05% transaction costs applied on each position adjustment. This is a parameter that can be adjusted in the backtest settings. Slippage, spread is not considered in the simplistic backtest that's on our site. We encourage using an independent backtesting infrastructure to validate performance.
When have been the factors developed? Is there an out-of-sample period?
Our cutoff date for our research process is 2024-01-01, we consider any data onwards as out-of-sample. Many factors have been traded live in some form from 2025-01-01.
Some of the factors look great! Should I pick an individual factor and start trading it?
We recommend using a portfolio of factors to maximize risk-adjusted returns. Alpha from individual factors is simply not consistent enough, and even running a simple arithmetic average of 3-5 relatively orthogonal factors will produce superior results.
Replication & API
Retail Flow Factor Analysis
Retail Flow factor analysis with AlphaLens
Replicate Spectra — our flagship multi-factor portfolio
Reproduce Spectra, our 6-factor market-neutral portfolio, from its constituent Unravel factors
Replicate Portfolio Backtest
Replicate a portfolio backtest with Unravel API
Get Live Weights
Get the live weights of the portfolio or the live factor data
