Stock
The history of a single stock: what the system said each day and what happened next. It is the most detailed screen in the application. At the top is the tracking chart, below it four tabs.
Header and controls
What it is: Ticker, name, sector, last price, daily change and the latest consensus state; below them the time range, horizon, tier selector and the "Show models separately" switch.
What data feeds it: /api/stock/{code}: the price_daily, return_daily, signal_daily, signal_outcome and event tables. The state badge is the latest D · Consensus signal at the selected horizon.
What it shows: Determines what the chart draws for the selected range and tier. The Live tier consists of records written before the outcome was known; Replay and Backtest are the model applied to the past.
What can be concluded from it: Turning on Live and Backtest together and comparing them is the quickest way to see side by side what the model promised in the past and what it actually did. The selections are kept in the address bar; whoever you share the link with sees the same view.
Tracking chart · top panel
What it is: The adjusted close line with direction arrows, hit lines, alarm bands and event marks drawn on it.
What data feeds it: price_daily (price), signal_daily (arrows), signal_outcome (hits), event (diamonds). In the default view only the consensus (D) arrows are drawn; the A and B arrows are turned on with the switch.
What it shows: An upward green triangle is an expected rise, a downward red triangle an expected fall; the size is the strength of the direction. Each arrow is connected with a thin line to the price at the end of the horizon. Amber bands are the days on which the activity of C · Change detection exceeded 0.5. Diamonds are recorded events. A hit is defined the same way everywhere: the signal's direction and the sign of the residual return at the end of the horizon are the same. The residual return is the stock's return with the effect of the market (XU100), the sector and USD/TRY removed; it is not the price itself.
What can be concluded from it: A line can be red even though the price rose: if the stock rose less than the market, its residual return is negative. Check whether the alarm bands come before or after large price moves; if they come after, the system catches the change but does not give advance warning.
Model and stock movement
What it is: A chart overlaying the selected model's signal each day and the stock's movement after the signal, and how often the model's direction was right in the selected range.
What data feeds it: /api/stock/{code}: signal_daily (the model's direction) and signal_outcome (the raw and residual return realised over the horizon after the signal, and the hit). The horizon and tier come from the selections at the top of the screen; if a day has more than one tier, one is used in the order live, replay, backtest.
What it shows: The bars are the stock's movement (percent; green same sign as the model's direction, red opposite, grey not counted), the line is the model's direction (-1 down, +1 up), and the bottom strip is the share correct among the last 20 resolved signals. At the top right: the share correct in the range selected with the slider, how many signals it was computed from, the 95% confidence interval, and the "if it had always said up" comparison.
What can be concluded from it: If the confidence interval of the share includes 50%, the result can be explained by chance. If the model does not beat the "always said up" share, it adds no directional information and merely reflects a rising period. Models are scored on residual return; a day that is right on raw return but wrong on residual is a day on which the stock rose but lagged the market. The sample for a single stock is small; for the models' overall hit rate see the Performance screen.
Tracking chart · middle panel
What it is: Daily residual return bars and the volume z-score line.
What data feeds it: The return_daily table: the residual and volume_z columns. The volume z-score is how many standard deviations the logarithm of daily volume is from its 60-day average.
What it shows: It shows the part of the stock's movement that the market, the sector and the exchange rate cannot explain, and whether trading volume is unusual.
What can be concluded from it: A long bar suggests stock-specific news or flow. A volume z-score above 2 means unusual interest. These are exactly the series that trigger the alarms of C · Change detection.
Tracking chart · bottom panel
What it is: The activity of C · Change detection (amber area), the models' direction series and, at the very bottom, the consensus state strip.
What data feeds it: signal_daily: the activity field of model C, the direction field of models A, B and D, and the state field of model D.
What it shows: Activity is the probability, between 0 and 1, that "something is happening now". Direction is between -1 and 1. Colours on the strip: green strong signal, amber event with unclear direction, grey weak, light grey none.
What can be concluded from it: Direction lines diverging while activity rises is the visual counterpart of "there is an event but it is unclear what will happen". Direction lines staying close to zero for a long time show that the model produces no information for this stock.
The investment information, comments and recommendations given here are not within the scope of investment advisory services. Investment advisory services are provided under an investment advisory agreement to be signed between a client and brokerage houses, portfolio management companies, or banks that do not accept deposits. The signals here are produced from historical data with statistical models, are shown the same to everyone and are not personalised; they may not suit your financial situation or your risk and return preferences. Therefore, making investment decisions based solely on the information given here may not produce results that meet your expectations.