Keeping up with a strategy

Golden Path Strategies is a research tool and never trades. This page shows what keeping up with a strategy looks like, from a bar closing to the alert that reports it.

Golden Path Strategies shows you what a set of fixed rules is doing and what it would have done in the past. It doesn't connect to an exchange or a broker and it can't place an order. This page is about keeping up with a strategy: when its signals change, how the site reports them, and why a backtest and real-world results drift apart.

None of this is advice. The site tells you what a strategy's rules say; every decision about them stays with you.

What Long and Flat mean

Every strategy is long-only, so it's always in one of two states.

  • Long means the rules want to own the market. For a BTCUSD strategy, the backtest account holds BTC (Bitcoin).
  • Flat means the rules have stepped aside, and the backtest account sits outside the market, in cash. A strategy never goes short, so Flat never means betting on a fall.

The backtest treats that cash as earning nothing, so any interest cash could earn in the real world isn't in the numbers.

When a bar closes and when the site updates

Every strategy decides once per 5-day bar, and only on bars that have finished. In late October 2026, for example, the crypto bars run 13 to 17 October, 18 to 22 October and 23 to 27 October; stock funds such as SOXX get roughly one bar a week. How the backtests are built explains how the bars are cut.

The site updates once a night, shortly after 00:10 UTC (about 1 to 2 am in Central Europe). It adds the new prices, brings every strategy up to date and then checks alerts against the bars that have just closed. A Long or Flat that appears in that update is final for that bar; a bar that's still forming is never used.

In the backtest, most strategies fill at the closing price of the bar that gave the signal; for crypto that's the same moment the next bar opens. The alert only goes out after the nightly update, so anything that happens the morning after an alert is already some hours behind the model. That's normal; how much it matters depends on how far the price moved in between.

The 5-day timeframe isn't a setting you can change; About explains why it was chosen.

A walk-through with Von Lohengramm T7

Take Von Lohengramm T7 - BTCUSD, one of the BTC strategies on the Home page. It's an index of seven strategies sharing nine votes; it buys BTC when five votes are bullish and steps aside when five turn flat. Over more than eleven years it made 16 trades, which means 32 orders, or about three a year.

Its two most recent trades, as a member sees them on the Trades tab:

[Trade 16: the entry on 2026-07-15 at about $64,700 and the exit on 2026-10-08 (2) at about $82,600, a gain of 27.6%.]

Here's how those two trades looked from the outside.

  1. Mid-July. The bar of 10 to 14 July closed with enough bullish votes. Shortly after 00:10 UTC on 15 July the site updated, the pill turned Long and an alert e-mail went out. The backtest bought at about $64,700. By the morning, the market price had already moved on from that close.
  2. August and September. Nothing happens. The strategy stays Long bar after bar and no e-mails arrive.
  3. Early October. The bar of 3 to 7 October closed with five flat votes. Shortly after midnight on 8 October the pill turned Flat and a second e-mail arrived. The backtest sold at about $82,600 and the account went back to cash.
  4. Since then. The strategy is Flat and waits for its next entry. That could be weeks or months away: after the October 2025 exit it stayed out until April 2026.

That's the whole rhythm: two signals in three months, each reported by an e-mail.

If you join while a strategy is already Long

Suppose you discover a strategy that's been Long for a while. Its open trade started at the entry price, not at today's price, so the gain it shows was made before you arrived. Starting to track it now and starting at its next entry give different pictures, and the strategy's statistics always describe its own trades, from its own entry points.

On the free version, the pill shows the position on the cutoff date, not today's. Von Lohengramm T7, for example, shows a visitor Long with the note "from 15 Jul 2026 until at least 28 Jul 2026, the end of the free version, about three months ago. It may have changed since." It did: it exited on 8 October. Never act on a free-version pill.

[The pill reads Long, with a note that this was the position at the end of the free version, about three months ago.]

All-in or all-out

A strategy is all-in or all-out of its own account. In the backtest, an entry puts the whole account into the market and an exit takes all of it out. Whatever is left after a trade, win or lose, is what the next trade starts with.

In a portfolio, each strategy works the same way inside its share. In Four Pillars Index [GPS], for example, crypto gets 50% of the portfolio, chip stocks 20%, BTC DCA 20% and gold 10%; inside crypto, half goes to BTC, so the BTC strategies together get 25%. The Trades tab of a portfolio lists amounts in an example account that starts at $100 and grows or shrinks with the portfolio, so read each amount against the account's size at that moment.

The drawdowns in the backtests are worth taking seriously: Von Lohengramm T7 once fell 42.9% from a high, even with its exits.

Let alerts do the watching

You don't need to check the site every day. With a plan, an e-mail tells you when something changes.

  1. Open the strategy and click the bell icon bell next to its name.
  2. The first time, type the address where alerts should go. It's saved to your profile.
  3. Under Description, add a note to yourself, for example why you're watching this strategy. The note is included in every e-mail, so future you doesn't have to remember.
  4. Click Create alert.

[The alert listed under Watching. 1. Test alert checks it against the latest finished bar. 2. Your note, which goes into each e-mail.]

All your alerts are listed on the Alerts page. Test alert (play icon) tells you whether the alert would fire on the latest bar, without sending anything. Following a whole portfolio? Put the bell on the portfolio and pick Alert Report: the e-mail then lists the portfolio's new allocation with one buy or sell line per asset. See Alerts for every option.

Why real-world results differ from the backtest

Real trading never matches the account line exactly, and it's worth knowing why.

  • Fees and spreads. For most strategies the backtest charges no trading fee at all, and a few charge a small one. Real exchanges and brokers charge their own fees, and real prices include a spread.
  • Timing. The backtest fills at the exact recorded price, for most strategies the close of the bar that gave the signal. A real order comes later, at a different price, sometimes a better one, often a worse one.
  • Taxes. In the real world each sale may be taxable. The backtest ignores tax entirely.
  • Stocks and gold. Some markets, such as the S&P 500 or gold, can't be bought directly at most brokers; funds and other products that track them behave slightly differently.
  • The future. A backtest is history. Rules that worked from 2015 to today can work less well, or not at all, from here on. See Risks and what this is not.

Golden Path Strategies provides analytics, not financial advice. It shows what fixed rules did and are doing; it doesn't tell anyone what to do.

Related