Crypto DCA Cost Basis Calculator

Purpose and model boundary

This calculator has two parts that share one set of inputs. Your purchases so far takes the purchases you have made (each with its date, the amount you paid and the price per unit) and the price today. It reports your units, your average cost, the value and the gain or loss at the price today, how much of that gain is on purchases held more than one year, and the gain on selling a share of your units under three orders of sale. A plan from today simulates buying a fixed amount every 1, 2 or 4 weeks for up to two years, starting at the price today, across 100 possible weekly price paths. The main result is your average cost; the page also shows the chance that the plan ends worth less than it put in.

To compare putting a sum in at once against spreading it out, use the Dollar Cost Averaging vs Lump Sum Calculator; for the tax on a gain, use the Capital Gains Tax Calculator. This calculator shows gains and holding periods, not the tax on them.

The model is educational. It is not a prediction of prices or tax law, and it is not financial, investment, tax or legal advice.

Inputs and units

Enter your purchases oldest first from the first row, up to 200, each with a date, an amount in dollars and a price per unit in dollars. Enter the price today and its date, which must be on or after your last purchase. The optional staking yield is a yearly percentage; blank means none. The share to sell is a share of the units you hold, so the compared sale can never exceed your holdings.

For the plan, enter the amount per purchase (at least 1 dollar), how often you buy, and the number of weeks (1 to 104). The market assumptions are the expected yearly return, the volatility of yearly returns, the shape of weekly returns (fat-tailed or normal) and, for fat tails, the tail heaviness from 3 to 30.

The workbook always runs 100 futures. Its optional seed defaults to 42 for the worked examples. On the page, Simulate uses the last result's next seed before a repeated run, while the first run after a worked example or Reset uses that example's seed. API and MCP callers can supply a seed to reproduce a run.

Governing relationships

A purchase of an amount a at a price p buys a / p units. Your average cost is the total amount over the total units, which is the amount-weighted harmonic mean of your prices. The value today is your units times the price today, and the gain or loss is that value less what you paid.

A purchase is held more than one year when the date of the price today is after the same day one year after the purchase, so one year to the day is still one year or less. In the United States a gain on a capital asset held more than one year is generally long-term; the holding period starts the day after the purchase and includes the day of sale. A purchase counts as in profit on its date when its price was above your running average cost including it; the first purchase equals its own average, so it never counts, and a relative tolerance of 1E-12 keeps rounding from deciding.

The compared sale takes whole purchases in a fixed order, then part of one: oldest first (FIFO), newest first (LIFO) or highest price first (HIFO; ties by date, then by row). The units taken from a purchase are the smaller of its units and what is still to sell after the purchases ahead of it. The gain is the units sold times the price today less the cost of the units taken, split into the part held one year or less and the part held longer. Selling the highest-cost units first gives the lowest gain of the three, but those units are often recent. US rules for digital assets generally identify the units sold by specific identification or first in, first out, within each wallet or account; LIFO and highest-cost-first are ways of specifically identifying units. The average-cost row is shown for comparison only.

The staking income is a yearly estimate at the price today: your units times the staking yield times the price today. Staking rewards are generally included in income when you gain control of them, at their fair market value then; the figure is not a tax amount, and the rewards are not added to your units.

For the plan, one seeded generator gives each future two consecutive parts of its stream, enough for 104 weeks. Each week's draw becomes a growth factor G = exp(m + s·Z), where m and s are the yearly lognormal pair matched to the expected return and volatility, divided by 52 and √52. For normal returns Z is a standard normal draw. For fat-tailed returns Z is a Student t draw with the chosen degrees of freedom, scaled to unit variance and capped at ±10 standard units. The cap keeps the rare extreme weeks finite; it lowers the standard deviation of the weekly draw by an amount that depends on the tail heaviness, about 4.32% at 3, 0.740% at 4, 0.17% at 5 and less above, so the weekly spread sits slightly below the entered volatility. With fat tails the median weekly growth is the same as with normal returns, and the simulated average weekly growth stays close to the one the expected return implies: within about 0.004% a week (about 0.2% a year) at the example's 70% volatility and tail heaviness 4, and within about 0.1% a week (up to about 4% a year) at the extremes of volatility and tail heaviness, where the expected return is only approximately the simulated average.

The plan buys at the start of week 0 and then every 1, 2 or 4 weeks before the last week, at that week's price. Its value after week w is V(w) = V(w − 1)·G(w) + a·f(w), where f(w) is 1 in a buying week. The units bought are the final value over the final price, and their average cost is the total put in over the units.

Calculation sequence

The workbook first checks the purchases and the inputs, then computes each purchase, the holding periods and the three orders of sale, then simulates the 100 futures and derives the weekly percentiles and the results. A repeated Simulate action uses the last result's next seed before the new run.

Outputs and interpretation

The cards show your average cost, the units you hold, the gain or loss at the price today, the staking income when you enter a yield, and the chance that the plan ends below what it put in, with its sampling error. The tables show every purchase, the compared sale by method, and the plan value by week.

The purchase chart spaces its columns by purchase, not by calendar time, and ends with a Today column. The plan chart shows the 10th, 25th, median, 75th and 90th percentiles by week and what the plan has put in so far. The chance of ending below what was put in comes from 100 futures, so it is reported with a sampling error of about ±9.8 percentage points.

Validation and status logic

The workbook evaluates the conditions below from top to bottom and returns the first matching status. This is the workbook's priority order, including warnings that may accompany a computed result.

NOT VALID is a refusal, not a numeric answer. CHECK may accompany a computable result but calls for the stated review. The self-check is separately reported as arithmetic-integrity evidence; it does not replace this ordered status chain.

Assumptions and limitations

Prices and values are nominal. The model computes no tax amount, and it does not model purchase or sale fees, a price history to buy along, staking rewards added to your units, your current units carried into the plan, choosing individual purchases to sell, wallets tracked separately, wash sales, or losses carried forward. The plan uses one expected return and one volatility throughout; beyond the fat tails, it has no mean reversion or change of market regime. The simulated prices are illustrations of the assumptions you enter, not forecasts.

Restrictions and non-computing states

This calculator enforces its numeric and choice limits before it accepts a result. The tail heaviness is read only when returns are fat-tailed, so a value entered for normal returns never turns a scenario into NOT VALID; a retained seed remains part of every simulation.

Errors and warnings

A request rejected by the page's input limits is not a computed result. A transport or service failure cannot establish a fresh workbook answer. Review the workbook's References and Disclaimer and current tax rules with a qualified professional before using a result for a real decision.

Sources and further reading

The model's purchases, prices and return assumptions come from the inputs, not market data. The workbook's References sheet identifies the calculation sources. The links below explain the rules and methods the model follows; they do not endorse this calculator.

• IRS Notice 2014-21 explains that virtual currency is treated as property for federal tax purposes, so a sale has a gain or loss measured against its basis (retrieved 2026-10-11).
• IRS Revenue Ruling 2023-14 explains that staking rewards are generally included in income when you gain control of them, at their fair market value then (retrieved 2026-10-11). The calculator's staking income is an estimate at the price today, not a tax amount.
• 26 U.S.C. § 1222 defines a long-term gain or loss as one on the sale or exchange of a capital asset held more than one year (retrieved 2026-10-11).
• IRS Publication 544, "Holding Period", explains the counting: start on the day after you acquired the property, and include the day you disposed of it, so an asset sold on the same date one year after it was bought has not been held longer than one year (retrieved 2026-10-11).
• 26 CFR § 1.1012-1(j), for digital assets acquired or disposed of on or after January 1, 2025: within each wallet or broker account, the units sold are those you specifically identify by the time of sale, otherwise the earliest acquired first (FIFO); identifying by earliest, latest or highest cost is not a method of accounting. Average basis under paragraph (e) is available only for fund shares and dividend-reinvestment-plan stock. The calculator's average-cost row is shown for comparison only (retrieved 2026-10-11).
• S. K. Park and K. W. Miller, “Random number generators: good ones are hard to find,” Communications of the ACM 31(10), 1988, and the Park–Miller–Stockmeyer 1993 correspondence describe the generator family. The workbook uses multiplier 48,271 and modulus 2³¹ − 1.
• The NIST/SEMATECH e-Handbook's lognormal and t distribution sections, and Microsoft Support's T.INV, NORM.S.INV and EDATE documentation, describe the distributions and date function the workbook uses.

This is an educational estimate, not financial, investment, tax or legal advice. Tax rules for digital assets change and depend on your situation; verify against current law and a qualified professional.