Tax-aware investing, explained from first principles.

How this is built

1. Built from the open literature

   The methods come from peer-reviewed research and the US tax code, worked out from first principles. Formulations in the code cite their public source (author, year and title), so their origin can be checked.

2. US tax rules in a small, tested library

   Wash sales (IRC §1091) and the short- versus long-term holding-period rule (§1222) live in a small lot-accounting library with no outside dependencies. A fixed set of worked test cases checks its behaviour, including partial and chained wash sales, the edges of the 30-day window, leap years and a replacement bought in an IRA.

3. An independent check on the accounting

   A separate validation harness, which imports none of the production code, rebuilds the lots, cost basis and short- and long-term gains from a backtest's trades. On a synthetic test run, checked on every change, it must agree with the backtester to the penny. It does not yet replay corporate actions or cash flows. A backtest can also run the harness's conservation checks after every rebalance: no cash or shares appear or vanish, lots add up to positions, and the optimizer's weights respect the budget, long-only and cash-buffer limits.

4. Reproducible runs

   The backtester is deterministic: the same code, data and settings produce the same trades, lots and returns, and an automated test confirms it on a synthetic run with every change. Each run also saves a manifest: a record of the code version, a fingerprint of each input dataset, the configuration and the solver settings.

5. Computed offline, served as static files

   Everything on the site was computed ahead of time and published as static files. There is no server, database or sign-in, and nothing is calculated in your browser.

6. Only derived numbers

   The site publishes only numbers derived from market data: no raw prices or other licensed data, and no licensed security identifiers. An automated test scans the published files for both.

7. Results only from real data

   Strategy pages show no performance numbers until backtests on real data are ready. Those results will be published with a forthcoming paper.

FAQ

Who is LotWise for?

Individuals, practitioners and advisors who want to see how tax-aware portfolios are built.

Is this investment advice?

No. LotWise is educational. Nothing on the site is investment, tax or legal advice, or a recommendation for your situation; the disclosures have the details.

Can I use it on my own portfolio?

No. LotWise is not a service: there are no accounts and nothing to upload. The tools run on sample portfolios, and the strategy pages explain the method rather than manage money.

Which strategies does it cover?

Two: tax-aware direct indexing and tax-aware long-short. Each strategy page sets out the objective, the constraints and the settings that change its behaviour.

Where are the backtest results?

A paper with results on real data is coming soon. The backtester simulates long-only portfolios and refuses a strategy that takes short positions, so there are no long-short results yet.

Are the tool numbers real?

They are real outputs of each tool's own code, run on curated sample portfolios, some tools at a few preset inputs. They show how a tool works, not your situation, and they price tax at top US federal rates, leaving out the 3.8% net investment income tax and state taxes.

Does the site collect personal data?

There is no sign-up, sign-in or form. The host measures aggregate page views and keeps standard request logs; the privacy policy has the details.

How do I get in touch?

By email: the contact page has the address.