About this site
This site replays what a split-buying rule would have done to a real
account, day by day, over historical closing prices. It is not a return
calculator. The point is to show the return and the damage on the same screen: how
deep the account fell while earning that return, and how often it sat underwater.
Why it exists
Leveraged ETFs are usually discussed in one sentence: "it went up N times in M
years." That sentence hides the 80% drawdown along the way, and it never asks whether
anyone could actually have held through it.
Rule-based approaches such as split buying make this worse, because the outcome
swings wildly on choices that sound minor: how many tranches, how far the price has to
fall before you buy, what profit target each tranche takes, and when you give up on a
tranche that is stuck. There was no easy way to see those differences side by side, so
we built a tool that takes the rule as numbers and replays the past with
it.
How the numbers are produced
These are the calculation rules, stated openly, because you need them before you can
decide whether to trust anything here.
- Whole shares only. No fractional investing. If a tranche is
allocated less than one share price, it buys nothing. That is why a small seed can
produce "no trades at all".
- Fill assumptions differ by family. Split buying (A) and dual mode
(D) fill both buys and sells at that day's close (limit-on-close). Breakout (B) and the
GDXU-only family (G) buy intraday at the breakout price and sell at the next open (or at
the stop). Infinite buying (I) checks its limit-on-close orders against the close and fills
them there, and fills the three-quarter limit sell when the day's high reaches it (at the open if the
stock opens above it); pre-market and after-hours fills are not modelled. Either way fills are assumed at exactly that price, so results are optimistic
by the slippage.
- Fees apply on both sides of every trade, 0.1% each way by default
and editable on screen. Taxes and slippage are not modelled, so the figures are
pre-tax and slightly optimistic.
- Maximum drawdown is measured on total account value including cash,
not on any single tranche's average cost. Split buying leaves cash idle most of the
time, so its drawdown looks smaller than buy and hold. That is not skill; it is half
your money sitting still.
- Every indicator uses the prior day's value only. Moving averages,
RSI, ADX and Treasury yields are never evaluated on the same day's close. Doing so would be lookahead, and the result would be a number nobody
could ever have achieved.
- Prices are split and dividend adjusted, so dividends are reflected
in the price. Taxes are not.
Data sources
- Prices — daily bars for US-listed ETFs and ETNs via
FinanceDataReader, loaded once a day into our own store rather than fetched per
request.
- US policy rate — the federal funds target from FRED at the Federal
Reserve Bank of St. Louis: the single target before December 2008 and the upper bound
of the target range afterwards, joined into one series.
We publish derived results computed on top of prices, not the prices themselves, and
there is no bulk download of raw market data.
What you can do here
- Backtest — pick a ticker, a period, a seed and a fee, then choose a
strategy. Strategies come in five families: split buying (A), dual mode (D),
overnight breakout (B), GDXU-only (G) and infinite buying (I), and every parameter is editable and savable. Compare up
to sixteen at once and export the thirty-year table as CSV or an image.
- Averaging-down simulation — add capital and start the same strategy
in a new account when the first one falls by a set amount, then see whether that
actually reduced the loss in windows that ended negative.
- Screener — which tickers meet the buy condition today.
- Today's orders — what prices your chosen rule produces for the next
session. It calculates; it does not place orders.
- Accounts — register a start date and capital, and see realised P&L,
monthly performance and current holdings as if you had followed the rule from then.
What this site does not do
It gives no investment advice and makes no recommendations. It is not connected to a
broker, places no orders, and never tells you to buy or sell anything. Every figure is
a simulation that assumes the rule was followed exactly on historical data, and nothing
guarantees the same rule produces the same result going forward.
We also state plainly that the strategies here were chosen after seeing how
they performed in the past. We do not hide that limitation: the guides show
which strategies lost their edge under a walk-forward test, where the choice is made
using only data available at the time.
Accounts and privacy
There is no sign-up. Saved strategies and account entries never leave your browser;
they live in local storage only. Clearing browser data removes them, and they do not
appear on another device. See the privacy policy for
details.
Who runs it, and how to reach us
This site is built and run by one person. It has no commercial sponsor, no
affiliation with any financial firm, and sells no products. Advertising covers the
running costs.
For bug reports, questions about how something is calculated, or requests for extra
tickers and strategies, write to
contact@backtestyourstrategy.com.
If a result looks wrong, please include the ticker, period, seed and
strategy so we can reproduce exactly what you saw.
Strategies · Privacy · Terms · Disclaimer · Contact
Change ad consent settings
Last updated: 2026-10-18