Open the simulator

The DCA calculator scoring rubric

8 axes, 24 points, each with four anchors. Fixed on 2026-08-26 — before any tool was opened — and used unchanged to score every calculator in the comparison.

This page exists because the comparison it feeds reviews our own product alongside twelve competitors. The only thing separating that from a rigged ranking is that the rules were written first, and the honest position is that you cannot verify that ordering: the repository holding commit 8402486 is private, so the hash is a reference rather than a check you can run.

What can be made public is this — the rubric itself, in full, rather than summarised. Every score in the comparison was assigned against these exact definitions, and the scorecard publishes all of them so you can hold any score against the anchor it claims.

The standard question each tool was asked to answer:

$50 a week into Bitcoin from 1 January 2018 to 1 January 2021. What is it worth, what did it cost, and how many bitcoin does it buy?

Most could not express it. That is the finding, and it is written up in the comparison.

The 8 axes and what each score means.
Axis0 · 1 · 2 · 3
Answers the question at allbacktest

Can you set an amount, a frequency and a start date, and get back what the plan would be worth, what it cost, and how much Bitcoin it bought?

  1. No historical backtest — a live price, a converter, or a forward projection only.
  2. Backtests, but from fixed presets; you cannot enter your own amount or frequency.
  3. Amount, frequency and start date are all yours to set.
  4. Also reports invested, current value AND accumulated units, not just a headline return.
Control of the windowwindow

Can you measure a closed period, or only "from a date until today"? A window you cannot close is a window you cannot cite, because the answer changes underneath the citation.

  1. Fixed window; no date control.
  2. Start date only, from a preset list.
  3. Any start date, always ending today.
  4. Any start date and any end date.
The same schedule, somewhere elsebenchmarks

Can you see what identical buys on identical dates would have done in something other than Bitcoin? A Bitcoin-only number tells you Bitcoin went up; it does not tell you whether the plan beat the alternative you actually had.

  1. Bitcoin only.
  2. Other cryptocurrencies only.
  3. At least one non-crypto benchmark (an index, gold, cash).
  4. Several, funded on the same schedule, comparable on one axis.
Where the prices came fromprovenance

Is the price source named on the site — which feed, how far back, how often it updates? Not "market data", but something a reader could go and check.

  1. Nothing. The numbers arrive from nowhere.
  2. A vague gesture — "live market data", an exchange logo, no specifics.
  3. The source is named.
  4. Named, with its start date and its update cadence.
The arithmetic, and what it leaves outmethod

Is it written down what happens on a weekend, how partial units round, and what the model does not include — fees, spreads, custody, tax? Every one of these tools ignores fees. The axis is not whether it ignores them; it is whether it tells you it does.

  1. No methodology, no exclusions stated.
  2. A one-line disclaimer, unrelated to the arithmetic.
  3. Either the arithmetic or the exclusions are disclosed.
  4. Both, in enough detail to reproduce the number by hand.
Can you keep the answer?permanence

Can you link someone to the exact run, or export it? A result you can only screenshot is a result you cannot cite, check later, or hand to anyone else.

  1. No. Reload and it is gone.
  2. State survives a reload on your own machine only.
  3. A shareable URL that reproduces the run.
  4. A shareable URL AND an export (CSV, image, or print) of the underlying rows.
Whose interest the tool servesindependence

Does the operator sell Bitcoin, custody it, or charge for the tool — and does the tool page say so where you can see it while reading the number? A calculator run by a seller is not disqualified. An undisclosed one is the problem.

  1. Operator has a direct commercial interest in the answer and the tool page does not say so.
  2. Commercial interest is disclosed, but only in a footer or a separate policy page.
  3. No commercial interest in the answer, or it is disclosed beside the result.
  4. Disclosed beside the result AND the tool works fully without buying anything.
What it costs you to get an answeraccess

Account? Email? Paywall? Does it work on a phone? Counted from a cold visit by someone who has never been to the site.

  1. Account or payment required before any result.
  2. Email or sign-up required to unlock the useful part.
  3. Full result with no account, but something material is gated.
  4. Full result, no account, nothing material gated, works on a phone.

Equal weighting is a choice, and it is ours: every axis is worth three points. A reader who cares most about getting an answer without an account, or about being able to send someone a link to a result, should weight access and permanence more heavily — and the published scorecard has the per-axis numbers to do that arithmetic.