All docs

Analysis

Read the metrics, filter trades, find better stop losses and save configurations.

The Analysis tab reads the trades saved for a strategy and turns them into charts, metrics and a trade table. It is where you decide which stop loss and take profit pair to keep, and which trades to filter out.

Trades come from manual backtesting and auto backtests. The page talks to the analysis server live. If the connection is not ready, the app says Analysis connection is not ready. Some analytics may need a plan that includes Advanced trade analytics.

Equity curve

The main chart draws one line for each stop loss and take profit pair. The title is Equity curve, or RR curve when you turn on Show RR curve.

  • Click a line to focus that pair. The metrics below describe the focused pair.
  • Stop loss and Take profit chip groups choose which pairs are drawn.
  • Include fees includes execution fees in the results.
  • Starting balance sets the balance the curve starts from.
  • Max expected drawdown (%) sets the drawdown the risk sizing aims for.

Key metrics

Until you focus a pair, the panel says No metrics yet. Click a line in the chart to focus a stop loss / take profit pair. After that it shows:

  • the pair, and whether fees are included,
  • Win rate, Wins, Losses, Total closed trades and breakeven trades. A trade that reached neither its stop loss nor its take profit counts as a loss of 1 R plus fees; the total then says how many, for example 53 still open, counted as losses,
  • Adjusted win rate,
  • Expectancy, observed and adjusted,
  • Average winning trade in R,
  • Risk sizing: Recommended risk and Adjusted full Kelly,
  • Net profit in R and as a percentage,
  • Annual projection,
  • Drawdown and streaks: maximum drawdown, Longest drawdown, actual and expected win and loss streaks,
  • Edge probability. The tooltip explains it is a model-based probability of a profitable sample, not the probability that the true edge is positive.

R is one unit of risk: a trade that loses its stop loss is -1R. "Adjusted" figures are the app's more cautious estimates. Treat them as estimates, not forecasts.

Test period and tried variants

Filters and Find best can always make the past look good: keep only the winning trades and any setup shows an edge. Two figures help you see through that.

Hold back a test period (the switch above the chart) splits the trades by start time. The older part is the research period; the newest part (30% by default, choose 10% to 50%) is held back. The line above the chart names both periods, for example Research period: 12 Mar 2024 to 3 Jun 2026 · 377 trades. Test period held back: 162 trades. While it is on, the chart, filters, calendar, trade list and Find best only use the research period. Key metrics adds In the test period: the same pair, filters and fee setting on the held-back trades, next to the research figures, with a verdict:

  • Holds up: positive per trade, an edge probability of at least 90% on the test trades alone, and not clearly worse than in research.
  • Not proven yet: positive, but too few test trades to rule out luck.
  • Weaker in the test period: positive, but further below the research figure than chance usually explains.
  • Does not hold up: zero or negative per trade in the test period.
  • Too few test trades to judge: fewer than 20 closed test trades after filters.

Every look at the test period with a different setup spends some of its value. The panel counts them. After more than 5 it says that the test period has effectively become part of the search. A test period needs at least 50 entered trades (30 for research, 20 for the test). Saved configurations remember the split; older ones load without it.

Edge probability after N variants. The edge probability assumes you picked one setup in advance. Every pair drawn on the chart, every applied filter change and every combination tested by Find best or a filter optimisation counts as a tried variant. Key metrics shows both, for example Edge probability 99.20% for a single pre-chosen setup; after 48 variants tried: at least 68.01%. The second figure is the chance that the best of 48 variants would not look this good by luck alone. Tried variants share trades, so the true figure is at least this high. Once more than one variant was tried, the verdict uses it. The count is kept per strategy; Start a new count resets it after you confirm.

Calendar and filter chart

  • Calendar shows results by Week, Month or Year, with the number of trades (Count).
  • Expectancy by filter value shows how expectancy changes across the values of a filter. Pick the filter in the Filter menu.

Filters

Choose Filters to restrict the trades that count. Number filters use sliders. Selection filters use chips. Indicator filters can use ranges.

  1. Adjust the filters you want.
  2. Choose Apply filters.
  3. Choose Reset to clear them.

Filters include the indicators you defined on the strategy.

Find the best settings

Find best at the top searches for the setting with the highest adjusted expectancy per trade. Choose one of:

  • Stop loss
  • Take profit
  • SL / TP combination

While it runs, the app shows how many combinations it has tested. When it finishes, a message names the best pair and its score.

Inside Filters, each filter has its own two buttons: Annual % optimises the filter for adjusted annualised return, and Expectancy % for adjusted expectancy per trade. Optimising on past data can overfit. A setting that looks best in the past may not hold up. Check that it still makes sense and that enough trades support it.

Save configurations

A configuration stores a stop loss and take profit pair together with your fee setting and filters, the test period split if one is on, and how many variants you had tried before saving.

  1. Focus the pair you want by clicking its line.
  2. Choose Save.
  3. Enter a Configuration name and choose Save.

The Saved configurations menu lists them. Choose one to apply it. Applying restores the pair, the fee setting and the filters. If the pair is no longer available, the app warns you that only fees and filters were applied. With a configuration selected, Save becomes Save changes (or Save as new). The trash button deletes a configuration.

Names must be unique. If a name is taken the dialog says so.

Your plan limits how many configurations you can save. See Plans and billing.

Saved configurations are used elsewhere:

  • The dashboard shows them on each strategy card, and the Portfolio list on the Edit tab chooses which count towards the combined balance.
  • AI training trains on a saved configuration.
  • Live automations run a saved configuration.

Trade table

Trades lists the trades behind the numbers.

  • Columns: Start Date, Trigger Date, Long / Short, Max R, Execution Fees, Pair Name, Price Points and Tags. Choose a column header to sort.
  • Sort order switches between Newest and Oldest.
  • Search tags filters by tag.
  • The Open trade chart button opens the trade on a chart so you can see how it played out.
  • The footer pages through results with First, Previous, Next and Last.