Dynamic Bitcoin DCA vs. standard DCA, across regimes.
Buys in tiers of 5x/10x/15x the base weekly amount as risk falls through the accumulate zone; never sells — an apples-to-apples comparison with weekly DCA, which also never sells. Same simulation shipped in the live DCA calculator. All figures are after capital-gains tax on a full liquidation and after trading fees.
Three kinds of evidence — kept separate on purpose
These are not the same claim, so we never blur them. Weakest-to-strongest, here's what stands behind the strategy.
Illustrative simulation
The regime comparison on this page. In-sample and illustrative — a demonstration of the strategy's logic on history it was calibrated against, not proof of future results.
You are hereReal money, since Dec 2021
The founder has traded this approach live, by hand, with real capital through a full cycle — buying the Dec-2021 top and holding through the 2022 drawdown. Private and unaudited, so treat it as background, not proof.
Read the story →Published & hash-chained
Every weekly call from both Anchor models, forward-tracked since the 2026-07-09weight freeze, append-only and verifiable. The strongest evidence — because you don't have to trust us.
See the record →These numbers are in-sample and illustrative — not a validated track record.
The model's factor weights were chosen by judgment and refined by looking at how the model behaved on these same historical cycles — they were not fixed prospectively before seeing the results. That means the backtests below should be read as a demonstration of the strategy's logic, not evidence it will perform this way going forward. See the methodology limitations and the full disclosure before drawing conclusions. For the longer argument, read why we call our backtests illustrative, not validated.
Public forward tracking started 2026-07-09. That is the date the live crypto composite's weights and zone bands were locked, specifically to stop in-sample tuning and begin an out-of-sample record. Everything on this page today is the pre-freeze backtest; every weekly call from 2026-07-09 onward is published on the live track record as the first evidence you can actually check of how the model performs on data it wasn't tuned against. The founder has traded this approach with real money since December 2021, but that record is private and by hand — it is context for where the model came from, not proof, and we don't count it here. See the methodology for why we keep those two apart.
Results by regime
Same $100 base weekly amount, weekly cadence, prior-week signal (no lookahead) for both strategies — the only difference is when and how much each buys. Because the two deploy different total amounts, the chart plots the value of $100 invested (return per dollar), not raw account balance — so the lines, the ROI, and the drawdown all describe the same thing. Select a window to see its curve.
All windows at a glance
| Window | Anchor Accumulate | Standard DCA | Alpha | Capital deployed |
|---|---|---|---|---|
| Full history: 2018 – present | +497.4% | +256.9% | +240.5pp | $87,000 vs $44,600 |
| Full cycle: 2018 bear → 2021 top → 2022 bear | +102.4% | +47.7% | +54.7pp | $67,000 vs $25,900 |
| Sustained bull: 2023–25 recovery | +206.8% | +67% | +139.7pp | $3,000 vs $15,600 |
| Bear / chop: 2021 top → 2022 bottom | -19.7% | -37.1% | +17.5pp | $32,000 vs $5,900 |
| 2021–present | +119.2% | +43.9% | +75.4pp | $52,000 vs $29,000 |
Alpha is return per dollar deployed, and the two rules deploy different amounts on purpose — Anchor Accumulate buys harder when risk is low and pauses when it isn't. A higher percentage therefore doesn't always mean more money. In the 2023–25 rally it deployed $3,000against Standard DCA's $15,600, so its stronger return per dollar still ended in a smaller position. Read both columns together. All figures are after capital-gains tax and trading fees.
Why this page shows the more conservative strategy
In backtesting, Anchor Bridge — the strategy that also trims into the S&P 500 — returned more than what you see here. We show Anchor Accumulate anyway, on purpose. A backtest's headline is the best a model found on the very history it was calibrated against, and leading with the biggest possible number would quietly set an expectation we can't stand behind.
A figure like +600% over a full cycle is not something to expect again — future returns can be far lower, or negative. We'd rather under-promise here than dangle a number that reads like a forecast. Anchor Bridge isn't hidden: it's shown where it actually counts — logged live, week by week, on the public track record, not dressed up in a backtest.
Run it with your own numbers.
The same simulation ships in the DCA calculator — plug in your amount and cadence.
Not financial advice. Backtest generated with model vriskModelComposite (18-factor, same live score as /crypto/risk-model), data as of 2026-07-26. Results are illustrative and in-sample; past performance does not predict future results. See the full disclosure.