Mansori

How it works

Every hour, a scanner downloads a year or more of daily prices for 145 stocks, ETFs and coins, runs the same six checks on each one, and publishes the result. The rules are simple on purpose, so you can see exactly why something is rated the way it is.

The six checks

  1. Trend. Is today's price above its 50-day and 200-day averages, and is the 50-day above the 200-day? Each yes scores a point, from 0 to 3.
  2. Momentum. The return over the past 12 months, leaving out the latest month because very recent moves often reverse. It is divided by how much the name usually swings, so a calm 20% gain counts for more than a wild one.
  3. Risk. How much the price usually moves in a year, measured over the last three months. Under 16% is Steady, under 32% is Balanced, and anything above is Bold.
  4. Timing. A pullback of 3% to 12% inside an uptrend, or a calm RSI between 38 and 56, is marked Good time to buy. An RSI above 66 is marked Ran up fast.
  5. Dip. Has the price dropped from its high of the last three months by more than about 40% of its usual yearly swing (at least 6%)? If so, the dip agent reads the completed daily closes. Still falling: a new low in the last three days, or below its 10-day average and lower than a week ago. Settling: no new low for a few days, no real bounce yet. Found a floor: at least four days above the low, back above its 10-day average, and up from the low. Bounced back: it has won back most of the drop.
  6. News. Recent headlines are checked against a short list of positive and negative words. If they lean clearly negative, the page says so. Headlines never change a rating on their own.

Buy, Wait and Pass

Buy needs a trend score of at least 2, strong momentum for its level of risk, and an RSI under 74. Only the 25 strongest names that pass get Buy, so the label stays meaningful when the whole market is rising. Pass means a weak trend and a falling year. Everything else is Wait.

The three lists

Every name sits in one list: Buy now, Wait or Pass. The price rules above decide which list. The news agent then groups each list by the main reason it found in the headlines from the past month: new products, partnerships and deals, government and politics, or analysts raising their views for Buys; still falling, economy worries or cautious analysts for Waits; misses, cuts, lawsuits or warnings for Passes. When no headline explains it, the reason comes from the price alone. A headline never moves a name into Buy by itself: exciting news often comes after the price has already moved.

Buying the dip

Many people like to buy after a price drops. The dip agent turns that into two simple rules. While a name is still falling, a Buy becomes Wait, even if its long-term trend is strong. When a name has found a floor and its 200-day average is still rising, it can become a Buy even if the trend score is low. That is a dip buy. At most 12 names get it, and they never take a spot from the 25 trend Buys. If the long-term trend is falling too, a floor is treated as a possible pause in a bigger slide, and the rating stays Wait.

Be clear about what this means. A bottom is only certain afterwards. "Found a floor" means the price stopped falling for now, not that it can't fall again. We tested these rules on the past year of prices for all 145 names. After removing the market's own move, neither waiting for a floor nor buying while it was still falling gave a reliable edge. In that year, names that had found a floor did slightly worse than average over the next four weeks. The rule is a way to avoid buying in the middle of a slide, not a way to beat the market. Dip buys are graded on their own on the scorecard, so you can see how they do.

On every pick you can drag across the chart to see how today's dip rule reads each past day, and what happened to the price after. This uses today's rule on old prices. It is not a record of what Mansori said back then.

Forecast ranges

Nobody can tell you where a price will be next month. Mansori shows a range instead. It starts from a modest long-run return (about 6% to 7% a year), nudges it slightly toward the past year's momentum, and spreads it by the name's usual swing over four weeks. Roughly two times in three, the price should end up inside that range. One time in three it won't.

Your money projection

When you record a purchase, Mansori values it every hour and projects a low, middle and high case for 12 months from now, using the same return and swing numbers and assuming your holdings partly move together. It is an illustration, not a promise. Your list stays in your browser on your device.

The scorecard

Each day the ratings are frozen and never edited. After four weeks, every Buy is graded against its market (the S&P 500 for US names, the TSX for Canadian ones, Bitcoin for coins). Losses will be shown as clearly as wins.

What Mansori can't do

Data sources

Prices: Yahoo Finance public charts. Headlines: Google News feeds. The news agent reads the headlines, says whether they sound upbeat or worried and what they are mostly about, and shows each headline with the publisher's name and date. Mansori does not copy articles.