Were we right?
Every directional call this site makes is written down at the moment it is made, with the level it was made at, and scored later against what actually happened. Hits and misses both. This page is the record.
Each mark below is one call. Above the line it made money, below the line it lost, and the taller the mark the more the move was worth.
How this is scored
A call is recorded when it is generated, together with the index level at that moment, and scored when its horizon passes — the next session's close for an Outlook lean, thirty minutes for a Scalper trigger. The level it is scored against is read back from the recorded tape at that horizon, not from the live market at whatever moment the scorer happens to run, so a verdict is a fact about the past and gives the same answer every time it is recomputed. Scoring only ever fills in the result on a row that already existed: no path in the code adds a call to the past, and none deletes one that went wrong.
A small move counts as neither a hit nor a miss. An index that closes 0.05% up has not vindicated a bullish lean, so moves inside a dead band sit on the centre line of the tape rather than counting as wins. Where an engine publishes its own expected move, that is used as the band — the market's idea of a normal day beats a constant.
"Wait" is not a win. When the Scalper stands aside it is making no directional claim, so those sit on the line too — counting them as hits is the oldest way to flatter a track record. How often standing aside was right, meaning the market then did nothing worth trading, is reported beside the record rather than inside it.
Past accuracy is a record of what happened, not a promise about what will. Educational only, not investment advice.
Why this page exists
Every site that publishes market readings should publish how those readings did. Almost none do, because the results are rarely flattering and nobody is obliged to show them. This is ours, kept whether or not it looks good.
What is being scored
The directional readings this site produces — the option chain's buildup bias, the intraday setups, the daily rankings — are recorded when they are made, with the level and the reasoning, and scored afterwards against what the market actually did. Calls are written down before the outcome is known, which is the only way a record means anything.
Nothing is added retrospectively and nothing is removed for being wrong. A record you can only see when it flatters the author is not a record; it is marketing with a table in it.
How to read a hit rate honestly
A hit rate on its own says very little. A method that is right 70% of the time and loses more on its misses than it makes on its hits loses money; one that is right 40% of the time with the opposite ratio makes it. The number to look for is not how often but how much.
Sample size matters more than most people allow for. Over a few dozen calls, the difference between a genuine edge and luck is not statistically visible, and neither a good run nor a bad one should change your view of the method much. Be as sceptical of a flattering stretch here as of an unflattering one.
What a good score would and wouldn't prove
Even a strong record here would not establish that these readings predict the market. They are fixed formulas applied to public data, they are not backtested against history, and market conditions change underneath any rule that isn't. A method that worked through a trending quarter can fail completely in a range.
The page is here so you can calibrate how much weight to put on anything this site prints — which should be some, and never all. Nothing on this site is investment advice; see the disclaimer.