Conditional odds · every number against its baseline
Describe a day like today. The board counts every day in history that looked the same and shows what happened next — beside what happens on an ordinary day, so you can see whether it mattered.
Pick a day like today. We find every day since 1993 that matched and look at what came next.
The big number is how often the market finished higher. The dashed amber tick on each bar is how often it finishes higher on any day at all. If the bar does not reach the tick, matching days were no better than ordinary ones.
The bar's colour is the odds themselves. A coin flip is grey. The further past half it goes the greener it runs, and the further under it falls the redder — so a bar tells you how sure a thing was before you have read a single number. Colour is not a verdict: whether the outcome beat an ordinary day is the amber tick the bar either reaches or does not, and the signed figure beside the number.
That tick is the whole point. “Up 62% of the time” sounds great until you learn the market is up 54% of the time anyway. Most places do not tell you the second number. We put it on every single one.
The median gain or loss is the other half. How often it went up does not tell you how far. The median is the middle one of the days it actually went that way — half were bigger, half smaller — and it has its own normal-day number beside it, same as everything else here.
Finished higher and finished lower are the same fact twice, quoted from both sides the way a board would. They do not quite add to 100%: a handful of days close at exactly the price they opened from, and those belong to neither. The two medians are not mirrors, though — when up days and down days are different sizes, that is where you see it.
The range in small type is how wrong the percentage could be, given how many days we found. Few days, wide range, less to trust.
This is a record of what happened, not a forecast. The market has no memory and owes you nothing.