All stocks
This tool is a decision aid, not financial advice. Every number here is an estimate based on past prices and published figures, and it can be wrong. Nothing can predict what a share price will do next. Decide for yourself, risk only what you can afford to lose, and consider talking to a registered adviser.
Guide to the Lagna Engine
This guide explains everything the app shows you, in ordinary words. You do not need any background in finance to read it. Start at the top and work down, or jump straight to the part you need.
If you read only one section, read Section 4. It is the one about money.
- Start here: what this tool actually does
- How to read the main table
- The verdict, and what to do about each one
- The trade plan numbers (the important one)
- The four scores and the weight sliders
- The pre-trade checklist
- Filters and screens
- Your lists: watching versus owning
- Where the data comes from
- Honest limits: what this cannot do
- A worked example from start to finish
- Common questions
- The Market mood tab
- The "Test one stock" tab, and what a backtest is
1. Start here: what this tool actually does
In one sentence: this app takes the list of companies you exported from screener.in, adds the daily price history of every stock listed on the National Stock Exchange, grades each stock out of 100, and tells you whether it currently looks like something to buy, something to wait on, or something to leave alone.
It does three things, in this order:
- It reads the price history. For every stock it knows the closing price on every trading day going back as far as it has records. From that it works out things like the average price over the last 50 days, how much the price bounces around day to day, and whether the stock has been going up or down.
- It reads the business numbers. These come from the spreadsheet files you download from screener.in. Profit growth, debt, how much the founders own, and so on.
- It combines the two into one answer. A score out of 100, a one-line verdict, and a set of actual rupee prices: where to buy, where to sell if you are wrong, and where to take profit if you are right.
Please read this part slowly
This is a decision aid, not advice. It is a calculator that does arithmetic you would find tedious to do by hand for two thousand stocks. It has no opinion, no inside knowledge and no crystal ball.
It will be wrong, regularly. A stock marked "Buy now" can fall the next day. A stock marked "Avoid" can double. Nobody, and no computer program, can predict what a share price will do. Everything here is a statement about the odds, based on what has happened before. The odds are not a promise.
Use it to narrow a huge list down to a handful worth a proper look. Then do that proper look yourself, and never invest money you cannot afford to lose.
2. How to read the main table
The big table at the bottom of the dashboard is where you spend most of your time. Each row is one company. Here is what every column means.
Symbol
The short trading code the exchange uses for the company. For example, the code for Tata Consultancy Services is TCS. This is what you type into your broker's app when you want to place an order.
Name
The full registered name of the company, so you know which business you are actually looking at. Several companies have confusingly similar short codes.
Sector
The industry the company works in: banking, cement, software, pharmaceuticals and so on. This matters because companies in the same sector tend to rise and fall together. If you already own three banks, buying a fourth bank is not really spreading your risk.
Price
The price of one share in rupees. While the market is open this is refreshed every few minutes, and the chip in the header says when it was last fetched. Outside market hours, and whenever live prices are unavailable, it is the closing price of the last full trading day that has been downloaded.
The two are not interchangeable, which is why the header tells you which you are looking at. A live price is the last trade that happened; a closing price is settled and is what every average, score and backtest in this app is calculated from. Either way, check your broker's screen before you place an order — by the time you act, the market has moved again.
The verdict badge
A coloured label with the app's one-line answer. It will say one of six things: "Buy now", "Good entry", "Buy on dips", "Watch, do not buy yet", "Good business, out of favour", or "Avoid". Section 3 explains each one and what to do about it.
Score
A number from 0 to 100. It is a blend of four separate grades: the price behaviour, the shape of the chart, the risk statistics, and the quality of the business. You control how much each of those four counts, using the sliders described in Section 5.
Rough reading: above 70 is strong, 50 to 70 is ordinary, below 40 is weak. A high score on its own is not a reason to buy. Check the verdict and the trade plan too.
Stop loss
A rupee price. If the stock falls to this level, the reason you bought it has stopped being true, and you sell and take the small loss. This is the single most useful number on the screen. Section 4 explains where it comes from and how to use it.
Target
A rupee price above today's price, where the app suggests taking some profit. There are two of them: a nearer one and a further one. Section 4 explains what to do at each.
Risk/reward
A ratio, shown as a single number such as 2.0. It compares how much you stand to lose if the stop loss is hit against how much you stand to gain if the first target is reached. A value of 2.0 means the possible gain is twice the possible loss. Bigger is better. Below about 1.5 the trade is not worth taking even if you are right about the direction, and Section 4 explains why.
Tags
Small labels showing which of the automatic screens this stock currently passes, for example "Momentum" or "Volume Breakout". They also show which of your downloaded spreadsheet files the company came from, which is handy if you keep separate exports for separate ideas. Section 7 explains every screen.
The small price line (sparkline)
The tiny squiggly line drawn inside the row. It is a miniature price chart of the recent past, with no numbers on it. It exists so you can glance down the table and see the general shape at once: rising, falling, flat, or jagged. It is a first impression, nothing more. A line that climbs steadily from left to right is a stock in an uptrend. A line that dives at the far right is a stock that has just fallen.
Opening the full chart
Click anywhere on the stock's symbol or name and a large chart opens at the top of the page, with the full price history and three panels of detail beside it: the business numbers, the price indicators, and the risk statistics. Use the Zoom In and Zoom Out buttons to look closer at a period, Reset Zoom to go back to the full view, and Close to put the chart away.
You can also click a column heading such as Score or Sector to sort the whole table by it. Click it again to sort the other way.
3. The verdict, and what to do about each one
Every stock gets exactly one of six verdicts. They run from most positive to most negative, except "Good business, out of favour", which is a different kind of answer and sits beside "Watch".
What it means. The app's strongest conviction (72 or more out of 100) in a stock whose long-term trend is up. The price may already be running, since in the past strong stocks have tended to keep going, but a stock that has shot far beyond its recent average (25 percent over the 20-day average, or extremely over-bought) is left out as a blow-off. Unlike the other verdicts it does not require a particular reward-to-risk ratio: that ratio is still shown, and still sizes your position, but it turned out to say little about what happens next. This is the app's highest rating, and a few dozen stocks can hold it on a given day.
What to do. If you want to act, this is the setup to act on. Buy inside the entry zone, place the stop loss order the same day, and size the position using the suggested percentage. Do not buy more than the suggested amount because the rating is high. A high rating is not the same as certainty.
What it means. One notch below "Buy now". Conviction is 68 or more out of 100 (stocks at 72 or more are normally "Buy now"), the long-term trend is up, and the stock is not over-bought (RSI at or below 72). Like "Buy now", it does not require a particular reward-to-risk ratio, because that ratio turned out to say little about what happens next. A stock that is over-bought is called "Buy on dips" instead.
What to do. Reasonable to buy, but look at the checklist first and decide whether the failing item bothers you. Consider taking a somewhat smaller position than the suggestion, since your confidence is lower.
What it means. The stock itself looks good, but the price today does not. It has gone up too much, too quickly. Either it is sitting far above its recent average price, or the momentum meter says buyers have been piling in unusually hard. Buying at this moment means paying a crowd price.
What to do. Do not buy today. Note the bottom of the entry zone, which is where the app thinks the price would be sensible again, and wait. Many stocks with this verdict will drift back down toward that level over the following days or weeks. Some will not, and you will miss them. That is the cost of not chasing, and it is a cost worth paying.
What it means. No hard stop applies, but the app is not convinced either, and there is no trigger. It is the largest group: roughly one stock in three. It is a mixed bag. In the app's own five-year replay, Watch stocks with a conviction of 60 or more did about as well as the average stock, while the rest did slightly worse over the following year, and those below their 200-day average did clearly worse. So "Watch" is not a quiet endorsement.
What to do. Nothing. If you like the company, add it to your Observation list (see Section 8) and let the app tell you when the verdict changes. Boredom is not a reason to buy something.
What it means. The other verdicts judge the chart. A strong company whose share price has fallen a long way has a poor chart, so it used to be called "Avoid", which reads as a verdict on the company when it is really a verdict on the timing. This label is for those companies. To qualify, a stock must pass every one of these: it is not a bank, lender or insurer; it is not a commodity producer (miners and metal makers look cheapest exactly when their commodity is at its peak); it is worth at least Rs 5,000 crore; its average return on equity has been at least 20 percent over both 3 and 5 years, and its return on capital is at least 20 percent; it owes no more than 0.3 times its equity; its profits and its sales have each grown at least 5 percent a year over three years; promoters have pledged at most 5 percent; its business score is at least 60; it trades enough to be bought and sold comfortably; its P/E is at least 25 percent below its industry's P/E; and its price is at least 25 percent below its highest of the last three-plus years (the app holds about five years of prices, so this is a five-year high, not a true all-time high; a company with under three years of prices is left out).
What it does not mean. The app cannot see a monopoly or a moat. A company that passes has the footprints a durable advantage usually leaves: years of high returns without borrowing, and growth. That is not proof. It also does not mean the price has stopped falling, and a price that has dropped a lot is not the same as cheap: "cheaper than its industry" is the only yardstick the data allows, and industry averages can be misleading. In the app's own five-year replay, the stocks that passed every test were few (11 companies) and did not beat the average stock over the next 3 to 12 months: typically they were down about 14 percent after a year, and up only 37 percent of the time, often because they kept falling. A broader group (good businesses that had crashed) did beat the average, but by clearly less than good businesses in an uptrend, and that comparison uses today's business scores, which flatters it. Five years of prices cannot test a bet meant to last longer, so this label asks for patience and research, not for optimism. The stop and target levels shown for these stocks are swing-trading levels and do not apply.
What to do. Treat it as a prompt to research the company, not as a signal to buy: why has the price fallen, is the advantage real, and are you willing to hold for years? Its own page lists which tests it passed. A safety blocker (thin trading, shares frozen at a price limit, heavy pledging, unpayable interest) still turns it into "Avoid".
About the "Business data" date in the header. The P/E, earnings and growth figures come from your Screener exports, which you refresh by hand, so unlike prices they can go stale; the date says how old they are and turns red after 45 days. Each time the exports change, the app also saves a snapshot of every company's valuation. After six snapshots a few weeks apart, a stock's page shows how its P/E compares with its own history (until then it says how many are saved), and a daily P/E line can be rebuilt from the saved earnings. Refresh the exports about once a month to build the history.
What it means. One of three things, and the stock's own page says which. (1) The numbers do not support buying. (2) The stock is in a downtrend or has just broken down: in the app's five-year replay, stocks in these two groups trailed the average stock by roughly 9 points over the following year. (3) The stock is hard to trade safely, for the reasons below. That third group is different: it is about whether you could buy and sell the stock, not a forecast, and those stocks did not fall more than average (many are small, thinly traded stocks whose past gains were not really available to a buyer). The serious problems that override everything else are: the stock is in a downtrend and below its average price of the last 200 days (unless it is a "Good business, out of favour"); the price has just broken down through a level that buyers were defending (same exception); less than about Rs 50 lakh a day changes hands, so you might not be able to sell when you want to; the price has been frozen at an exchange price limit on several recent days; the founders have pledged more than 40 percent of their shares as security against loans; the company is not earning enough to comfortably pay the interest on its debt; or the share costs less than 10 rupees, where the risk of manipulation is high.
What to do. Leave it. If you already own it, read the Active Positions advice in Section 8, which is a separate question from whether to buy more.
Conviction: a number from 0 to 100
Conviction is how strongly the evidence supports the verdict. It is not a prediction of how much money you will make. It is closer to "how many of my tests agreed with each other".
It is built like this. 30 points come from how many of the eight checklist items in Section 6 passed. Then 15 points come from the price-behaviour score, 15 from the chart-shape score, 25 from the business-quality score, and 15 from the risk statistics. Nothing is added on top: several of the checklist items and the chart scores look at the same price evidence, and counting it again as a "bonus" would let one good-looking chart count three times. The business has the biggest single share of the scores because nothing else measures it. Finally, if the app has less than a full year of price history for the stock, the whole conviction number is reduced by 15 percent, because a short record is a weaker piece of evidence.
Confidence: High, Medium or Low
Confidence is simply conviction translated into a word, so you do not have to remember what the numbers mean:
- High means conviction is 72 or above. Most of the evidence agrees.
- Medium means conviction is between 55 and 71. The evidence is mixed but leaning positive.
- Low means conviction is below 55. The evidence is thin or contradictory. Treat anything here with real caution, whatever the verdict says.
A useful habit: if Confidence is Low, do not act, no matter how much you like the company.
4. The trade plan numbers
This is the section that decides whether you make or lose money. The verdict tells you what the app thinks. The trade plan tells you exactly what to do with your rupees. Read it twice.
Our example stock for this whole section
Imagine a made-up company called Bharat Widgets Limited. It is not real. Today its share costs Rs 500.
Two more facts we will need. First, on an ordinary day this share moves about Rs 12 between its high and its low. That everyday movement has a name: the average daily range. Second, over the last few months the price has repeatedly fallen to around Rs 482 and then bounced back up. A level that buyers keep defending like that is called a support level.
Entry zone: the price range where buying makes sense
The entry zone is two prices, a low one and a high one. Buying anywhere between them is fine. Buying above the top of the zone is not.
The bottom of the zone is anchored near the stock's 20-day average price, which is simply the average of its closing prices over the last twenty trading days. That level is where a stock in an uptrend usually pulls back to when it takes a breather. The top of the zone is a fraction above today's price, so you are never told to chase a stock that has already jumped.
For Bharat Widgets, the entry zone is about Rs 488 to Rs 505.
Why buying above the zone matters so much. Your stop loss is a fixed price. It does not move up just because you paid more. Suppose you were patient and bought at Rs 490 with a stop at Rs 475. You are risking Rs 15 a share, about 3 percent. Now suppose you were impatient and bought the same stock at Rs 540 after a jump, with the same stop at Rs 475. You are now risking Rs 65 a share, about 12 percent, to chase the same target. The company has not changed. The idea has not changed. Only your price has changed, and it has turned a good bet into a bad one. This is the single most common way that sensible people lose money: not by picking the wrong stock, but by paying the wrong price for the right stock.
Stop loss: the price at which you admit you were wrong
A stop loss is a price you decide on before you buy. If the share falls to it, you sell, accept the small loss, and move on. You do not renegotiate with yourself. Most brokers let you place a stop-loss order at the same time as your buy order, so it happens automatically and you are not tempted to argue with it at three in the afternoon.
The app works out two candidate stop prices and then picks the safer one.
- The structural level. This sits just under the support level, the price where buyers have been stepping in. For Bharat Widgets, support is Rs 482. The app places the structural stop about one and a half percent below it, at roughly Rs 475. The logic: if the price slices through the floor that buyers were defending, the reason you bought has broken.
- The volatility level. This sits two times the average daily range below today's price. Bharat Widgets moves about Rs 12 on an ordinary day, so two times that is Rs 24, giving a volatility stop of Rs 476. The logic: your stop must be far enough away that normal, meaningless daily wobble does not knock you out of a perfectly healthy position.
Bharat Widgets ends up with a stop at Rs 475, which is 25 rupees, or 5 percent, below the current price.
Two safety rails are applied on top. The stop is never allowed to be more than 15 percent below the price, because no single short-term position should be able to hurt you that badly. And if the arithmetic somehow produces a stop above the current price, it is reset to 7 percent below instead.
A stop loss you do not honour is not a stop loss. It is a wish. Every number in this trade plan, including the position size, assumes you will actually sell when the price is hit.
Target 1 and Target 2: where to take profit
Deciding when to sell a winner is harder than deciding when to sell a loser, because there is no pain forcing your hand. So the app fixes two exit prices in advance.
Target 1 is the nearer one. It is set at twice your risk above the buy price. Bharat Widgets risks Rs 25 a share, so Target 1 is Rs 500 plus Rs 50, which is Rs 550, a gain of 10 percent. If there is a known ceiling in the way, a price the stock has repeatedly failed to get past, Target 1 is moved down to just below that ceiling instead. There is no sense aiming through a wall.
Target 2 is the further one, set at three and a half times your risk above the buy price, or further if the stock's own history says a bigger move is plausible. For Bharat Widgets that is Rs 500 plus Rs 87.50, which is Rs 588, a gain of about 17.5 percent.
Sell half at Target 1, then move your stop up
This is the most useful habit in the whole app, and it works like this.
You bought 200 shares of Bharat Widgets at Rs 500. The price reaches Target 1 at Rs 550. You sell 100 shares, banking Rs 5,000 of profit that nobody can take away from you. Then you raise your stop loss on the remaining 100 shares from Rs 475 up to Rs 500, which is exactly what you paid.
Look at what you have done. If the stock now collapses, your remaining shares sell at Rs 500 and you lose nothing on them, while keeping the Rs 5,000 you already banked. The trade can no longer lose you money. If instead the stock keeps climbing to Target 2 at Rs 588, you sell the rest there for another Rs 8,800. You have removed the downside and kept the upside. That trade-off is worth far more than the extra profit you occasionally give up by selling half early.
Risk and reward: the ratio that decides whether a trade is worth taking
Take the distance up to your target and divide it by the distance down to your stop. That is the risk/reward ratio.
The arithmetic. Suppose your stop is 5 percent below your buy price and your target is 12 percent above it. Divide 12 by 5 and you get 2.4. Your ratio is about 2.4 to 1: you are risking one rupee to try to make two rupees and forty paise.
Why this matters more than being right. Imagine you take ten trades at 2.4 to 1, risking Rs 10,000 each time, and you are only right four times out of ten. Your four winners make 4 times Rs 24,000, which is Rs 96,000. Your six losers cost 6 times Rs 10,000, which is Rs 60,000. You were wrong more often than you were right, and you still finished Rs 36,000 ahead.
Now run the same ten trades at a ratio of 1.2 to 1. Four winners make 4 times Rs 12,000, which is Rs 48,000. Six losers still cost Rs 60,000. Same skill, same hit rate, and you have lost Rs 12,000.
This is why anything under about 1.5 to 1 is a poor bet even when you are right about the direction. At a low ratio you need to be right most of the time just to break even, and nobody is right most of the time. The app's checklist only ticks the reward-versus-risk box when the ratio reaches 1.8 to 1.
A note on the numbers you will see: Target 1 is set at twice the risk, or closer if a ceiling is in the way. So the ratio measured at Target 1 will be 2.0 at best, and often a little less. Target 2 is further out and carries a higher ratio, but it is also reached less often.
Suggested position size: how much money to put in
People agonise over which stock to buy and then decide how much to buy by feel. That is backwards. How much you buy is what actually controls whether a bad run damages you.
The rule here is simple: size every position so that if the stop loss is hit, you lose about 1 percent of your total portfolio. Not 1 percent of the position. One percent of everything you have invested.
Here is the arithmetic for Bharat Widgets, assuming your whole portfolio is Rs 10,00,000.
- One percent of Rs 10,00,000 is Rs 10,000. That is the most this idea is allowed to cost you.
- Your stop is Rs 25 below your buy price, so each share can cost you at most Rs 25.
- Rs 10,000 divided by Rs 25 gives 400 shares.
- 400 shares at Rs 500 each is Rs 2,00,000, which is 20 percent of your portfolio.
- The app never suggests putting more than 10 percent of the portfolio into one stock, whatever the arithmetic says. So the suggestion is capped at 10 percent, which is Rs 1,00,000, or 200 shares. If the stop is hit you lose Rs 5,000, half a percent of your money.
Notice what falls out of this rule automatically. A jumpy stock needs a wide stop, so the arithmetic gives you fewer shares and a smaller position. A calm stock with a tight stop lets you hold more. You do not have to think about it. The maths sizes the risky ones down for you.
Two warnings. The 1 percent figure only works if you actually sell at the stop. And it assumes the price does not jump straight past your stop overnight, which occasionally happens. See the question about overnight gaps in Section 12.
5. The four scores and the weight sliders
The single score out of 100 in the table is a blend of four separate grades, each also out of 100. Each one looks at the stock from a completely different angle. A stock can be excellent on one and dreadful on another, which is exactly why they are kept apart.
Technical score: what the price is doing right now
This grade asks a simple question: is the price rising, and is it rising in a healthy way? It compares today's price with the average price over the last 20, 50 and 200 days. It checks a momentum meter that shows whether buyers have been more eager than sellers lately. And it checks whether the price has stretched unusually far from its own recent average, because a rubber band pulled too tight tends to snap back. A high score means a stock moving up steadily with room left to run. A low score means a stock falling, or one that has already gone vertical.
Pattern score: the shape the chart is drawing
Prices do not travel in straight lines. They zigzag between a floor where buyers keep stepping in, called support, and a ceiling where sellers keep appearing, called resistance. This grade works out where those two levels are and where the price currently sits between them. Sitting near the floor is a better place to buy than sitting under the ceiling, because the floor is close by if you are wrong. It also measures how steeply the price has been trending and how far below its all-time peak it is. A high score means a stock in a tidy uptrend, resting near its floor or breaking cleanly through its ceiling. A low score means a broken chart.
Statistical score: how rough the ride is
This grade ignores whether the stock is a good business or a pretty chart and asks only one thing: are you being paid enough for the stress of holding it? It measures how violently the price swings, compares the stock's returns against what you could get from a safe fixed deposit, checks whether it has done better or worse than the wider Indian market, and checks that enough shares trade each day for you to be able to sell when you want to. A high score means steady gains without white-knuckle swings. A low score means a wild ride for returns you could have got more calmly elsewhere.
Fundamental score: is this actually a good business?
This grade forgets the chart entirely and reads the company's accounts, using the figures from your screener.in export. Are sales and profits growing? How much money has the company borrowed, and can it comfortably pay the interest? Does it earn a good return on the money invested in it? Do the founders still own a large stake, and have they pledged their shares against loans? Is the share price reasonable compared with the profits the company makes? A high score means a growing, profitable, lightly indebted business. A low score means one that is shrinking, drowning in debt, or wildly expensive.
Also worth reading: stops and position sizing
Not a score, but the mechanics behind the stop loss and the suggested position size, explained in more detail than Section 4 has room for.
The weight sliders: telling the app what you care about
Under the four score names on the dashboard are four sliders. They decide how much each grade counts toward the single total score. There are two rules to know.
- The four weights always add up to 100. They are shares of one pie, not four independent dials.
- Moving one slider takes points from the others. If you push the business-quality weight up from 10 to 40, those extra 30 points have to be taken from the other three, and the app takes them proportionally. This is not a bug. It is what keeps the total honest.
The lock buttons
Beside each slider is a small padlock. Click it to lock that weight in place. A locked weight will not be touched when you move a different slider, so the adjustment is shared out among the remaining unlocked ones instead.
This is genuinely useful. Suppose you have decided that business quality must always be worth exactly 30 points, no matter what. Set it to 30, lock it, and then experiment freely with the other three. Click the padlock again to unlock it. If you lock so many sliders that there is nothing left to adjust, the remaining slider simply will not move, which is the correct behaviour.
The three presets
Above the sliders are three buttons that set all four weights at once. Each one describes a different way of investing and a different holding period.
Momentum Swing, for holding one to four weeks
Sets the weights to 40 technical, 40 pattern, 10 statistical, 10 fundamental. Eighty percent of the score now comes from the chart. Over a few weeks, what a company earned last year barely matters. What matters is whether the price is moving and whether the shape of the chart supports it continuing. Suits you if you check the market most days, you are comfortable selling quickly when a stop is hit, and you are trying to catch a move rather than own a business.
Position Trader, for holding three to twelve months
Sets the weights to 25 technical, 25 pattern, 20 statistical, 30 fundamental. A balanced setting. Over several months both the chart and the accounts have time to matter, and the risk statistics get a real say because you will be living with the swings for a while. Suits you if you look at your portfolio weekly rather than daily and want the business to be sound as well as the price to be moving. This is the sensible default for most people.
Value Investor, for holding one to five years
Sets the weights to 10 technical, 10 pattern, 20 statistical, 60 fundamental. Now the accounts dominate and the chart is almost ignored. Over five years a business that keeps growing its profits will do well regardless of what its chart looked like on the day you bought. Suits you if you buy a few companies, hold them for years, and do not want to be disturbed by every wobble. Be aware that with this setting the app will happily give a high score to a company whose price has been falling for a year.
You can start from a preset and then nudge the individual sliders. Nothing is locked in, and changing the weights re-scores and re-sorts the whole table immediately.
6. The pre-trade checklist
Every stock is run through the same eight questions. Each one gets a pass or a fail, and each shows the actual figure behind it. How many pass is the biggest single piece of conviction, so this list is worth understanding.
You will rarely see all eight pass at once. Five or six is a good setup.
1. Long-term trend is up
The price today is above its average price over the last 200 trading days, which is roughly the last year. This is the broadest possible test of whether the stock is generally going up or generally going down.
2. Medium-term trend is up
The same test over a shorter window: the price today is above its average price over the last 50 trading days, which is roughly the last ten weeks.
3. Momentum is turning up
A meter that compares recent price movement against slightly older price movement has just tipped positive, which suggests buyers are becoming more eager than they were. Traders call this meter the MACD, and this check passes when its faster line rises above its slower line.
4. Not overbought or oversold
A 0 to 100 speedometer of buying pressure, called the RSI, is sitting in a healthy middle range between 40 and 68. Very high means the crowd has piled in and a pullback is common. Very low means panic selling. Neither is a comfortable place to start.
5. Not stretched away from its average
The price is less than 12 percent above its average price over the last 20 days. Further than that and the stock has run up too fast, and buying now usually means buying just before it pauses to catch its breath.
6. Beating the market
Over the past year this stock has gone up more, or fallen less, than the NIFTY 500 index, which tracks 500 of India's largest listed companies. If a stock cannot beat the market, owning the market instead is simpler and cheaper.
7. Enough daily trading volume
At least 100,000 shares change hands on an average day. This is about being able to get out. In a thinly traded stock your own sell order can push the price down, and on a bad day there may be nobody on the other side at all.
8. Reward outweighs the risk
The distance up to the first target is at least 1.8 times the distance down to the stop loss. Section 4 explains why this ratio decides whether a trade is worth taking at all.
7. Filters and screens
Just above the table is one slim toolbar for cutting a list of thousands down to a list you can actually read. The things you reach for most are always visible; the rest sit in three drawers that stay closed until you open them, so the table stays near the top of the page.
The search box and sector
Type part of a company name or trading code and the table narrows as you type. The sector menu next to it shows only companies from one industry, which is useful when you want to compare like with like, or avoid yet another company in a sector you already own too much of. Choose "All sectors" to remove it.
Quick screens
The row of tick-able chips, shown to approved accounts only. Each is a specific pattern the app looks for, and the number beside it is how many stocks match today. Tick one and only stocks matching it remain.
Below fair value
Profitable companies whose share price is at least 30 percent below a cautious estimate of what the business is worth, calculated from earnings and three-year profit growth with a formula from the investor Benjamin Graham. Growth is capped at 10 percent a year, and the screen refuses to give a verdict, rather than guess, when earnings or growth are missing, when profit is falling, or when the figures do not agree with the current price (for example after a share split). It ignores the chart completely. Some days it may show very few stocks, or none. That is the screen being careful, not broken, and the ones it does show may still be out of favour for good reason.
Sound and growing
A business with a good fundamentals score built on enough measured data, debt of no more than the shareholders own money, comfortable interest cover, little pledged by the promoters, and profits that have grown over three years. It must also be above its 200-day average, not overbought, and traded in meaningful rupee volume. Banks, lenders and insurers are left out because debt means something different for them. Any missing figure counts as a fail, so this is the strictest screen and is often short.
Good risk and reward
Setups where the distance up to the first target is at least twice the distance down to the stop loss, excluding anything the plan marks Avoid. Section 4 explains why that ratio decides whether a trade is worth taking.
Breaking out
Stocks that have just pushed up out of a range, measured against the highs and lows from before today, on at least one and a half times the usual volume, and stocks making a new high on much heavier trading than usual. Thinly traded stocks are excluded. Only moves to the upside count: a stock falling out of its range is a breakdown, not a breakout, and is not shown here. Volume is the evidence that real money is behind the move.
Momentum
Stocks in a statistically reliable uptrend. The app fits a straight line through the last 20 closes and keeps the stock only if the slope is steep and also clearly more than noise (a t-statistic of at least 3), the price is above its 50-day average and that is above its 200-day average, and enough rupees change hands daily. Short price histories never qualify. The thinking is that a stock already moving tends to keep moving for a while, but a screen like this is late by nature: it finds trends that are already well under way.
Important: ticking several screens narrows, it does not widen
If you tick Momentum and Breaking out together, you get only the stocks that pass both tests. Not the stocks that pass either one. Every extra chip you tick makes the list smaller, and it is very easy to tick several and get an empty table. If your results vanish, untick one. Clear all removes every screen and filter at once.
The three drawers
- Filters. The Screener-file menu (every spreadsheet you put in the downloads folder becomes an option, so you can look at one list at a time), company size, rows per page, an "only stocks with Screener data" box, and three range sliders: Score (approved accounts), Price in rupees, and Market cap. Market cap runs from a few crore to lakhs of crore, so the slider is stretched to keep small companies usable. A small number on the button says how many of these filters are in use.
- Query. Build your own screen: pick a field (dividend yield, P/E, ROE and about fifty more), a comparison and a value, and add as many conditions as you like, joined by "all" or "any". The button shows how many stocks matched.
- Score weights. Approved accounts only. Sets how much each of the four underlying scores counts towards the overall score. See Section 5.
All of these work together. Search text, sector, every ticked screen, the filters and the query are applied at the same time, and a stock has to satisfy all of them to appear.
8. Your lists: watching versus owning
There are two personal lists, and the difference between them is whether your money is actually at risk.
The Watch tab: stocks you are watching
No money committed. This is your shortlist. When you see something interesting but the verdict says to wait, put it here so you do not have to find it again.
The app records the price on the day you added it, so you can see later whether waiting helped or cost you. It also shows the current price, the support level that would act as a stop if you did buy, and the ceiling above. Nothing here affects any calculation. It is a notepad that keeps itself up to date.
Add a stock with the observation button on its row, and remove it with the delete button on the list.
Active Positions: stocks you actually own
Real money. When you have actually bought a stock, press "I Bought This" on its row. The app records the price you paid, and from then on it can tell you something it could not before: whether you specifically are up or down, and by how much.
This changes the question entirely. For a stock you do not own, the question is "should I buy this?" For a stock you own, the question is "should I hold, add, trim, or get out?" Those have different answers, and the app keeps them separate.
Two separate opinions on the same stock
For every position you own, the app gives two independent recommendations, side by side.
Driven purely by price levels, and nothing else. The stop and the targets are measured from the price you paid, using the same rules the app uses for a new buyer today (a stop about two typical daily ranges below, or just under support, and a first target at twice that distance above), so the advice can tell you when a loss has gone as far as it should or a target has been reached. It will tell you to sell immediately if the price has fallen through your stop; to sell everything if the second target has been reached; to sell half and raise your stop to your purchase price if the first target has been reached; to consider trimming if the buying-pressure meter has gone extreme; or to sell if the chart has broken down through support on heavy volume. Otherwise it says hold, and reminds you where the stop and the next target sit.
Driven purely by the business and what you are paying for it, ignoring the chart entirely. It will tell you to sell if the business quality score has collapsed; to trim if the price has risen far above a rough estimate of fair value; to buy more if the price is well below fair value and the business still holds up; and otherwise to hold and leave it alone.
Why the same stock can be a sell for one and a hold for the other
This is not the app contradicting itself. It is the app refusing to pretend that two different questions have the same answer.
Take a genuinely excellent company whose price has just jumped 30 percent in a month on some good news. The short-term view says trim: the price has run far ahead of itself, the buying-pressure meter is at an extreme, and pullbacks from here are common. That is a correct reading of the next few weeks.
The multi-year view says hold: profits are compounding nicely, there is little debt, the founders own most of it, and none of that has changed because of one good month. That is also a correct reading of the next few years.
You decide which question you were asking when you bought. If you bought it as a trade, follow the trade view. If you bought it as a long-term holding, follow the long-term view. What you must not do is switch from one to the other whenever the answer is more comfortable. Deciding you are a long-term investor at the exact moment your trade goes wrong is the most expensive habit in the market, and having both views on screen makes it very easy to fall into.
9. Where the data comes from
Two sources, kept completely separate.
Prices: the exchange's own daily file
At the end of every trading day the National Stock Exchange publishes one official file containing the open, high, low, close and traded volume of every single security that traded that day. It is called the Bhavcopy. This app downloads that one file once per trading day and stores it on this computer.
Two consequences follow, and they are the reason it works this way:
- Every listed stock is covered. Not the popular few hundred. All of them. There is no company that is missing because some website did not have a page for it.
- Nothing gets rate-limited or blocked. The app makes one download per day rather than thousands of individual requests, so there is no throttling, no failed lookups, no half-loaded tables and no dependence on a service that might stop working. After the first download, everything is read from this computer's own disk, which is why the app is fast.
Because the file comes from the exchange itself, it is the same data your broker sees. There is no third party in the middle to introduce errors.
Adjusting for splits and bonus shares
Sometimes a company turns one share worth Rs 2,000 into ten shares worth Rs 200 each. This is called a split, and nothing of value has changed. A bonus issue does something similar by handing existing holders extra shares. On the chart, though, both look like the price suddenly collapsing by 90 percent.
If that were left alone, every average price, every trend and every risk figure would be nonsense for months afterwards. So the app detects these events automatically, by comparing the exchange's own stated previous close against the actual previous close, and rescales all the older prices to match the new share count. Volumes are rescaled the other way. You do not need to do anything about splits or bonus issues. The history you see is already corrected.
Business numbers: your screener.in exports
Prices tell you nothing about whether a company makes money. That comes from screener.in, a website where you can build a list of companies and download it as a spreadsheet.
Put those downloaded CSV files into the folder named downloads next to the app, and it reads every one of them when it starts. You can have as many files as you like. Each becomes a selectable option in the source-file dropdown described in Section 7. Companies are matched to their price history using the NSE Code column, so your export must include that column or the rows cannot be connected to anything.
Missing columns are skipped, never guessed
This is an important design decision. If your export does not include, say, the debt column, the app does not assume the company has no debt, and it does not assume an average amount of debt. Both of those would be inventing a fact.
Instead that part of the business score is simply left out, and the remaining parts are rescaled so the score is still out of 100. A company is never rewarded or punished for a column your spreadsheet happened not to contain.
Two things follow. First, when a lot of columns are missing the app pulls the score back toward a neutral 50, because it genuinely does not know enough to have a strong opinion. Second, the app tells you exactly which columns to add, by their precise screener.in names, so you can go back and re-export a more complete file. Adding the missing columns is the single easiest way to improve the quality of everything the app tells you.
Column names are matched forgivingly, so differences in capitalisation, spacing or punctuation between exports will not break anything.
10. Honest limits: what this cannot do
Every one of these is a real limitation, not a disclaimer written by a lawyer. Knowing them is the difference between using the tool well and being misled by it.
It only sees daily closing prices
The app knows what each stock opened at, its high, its low and its close for each day. It knows nothing about what happened between those points. A stock that crashed 8 percent at eleven in the morning and recovered by three o'clock looks, to this app, like a perfectly quiet day. Every signal is calculated once per day, from one day's numbers.
It cannot read the news
A regulatory investigation, a factory fire, a resignation, a lost contract, a change in government policy: the app knows none of it. It will only find out indirectly, days later, when the price and the next export reflect the damage. Before you buy anything, spend two minutes searching for recent news about the company.
It cannot judge management or an industry turning
Whether the people running the company are honest and capable is the single biggest factor in a long-term investment, and no number in any spreadsheet captures it. The app uses crude proxies, such as how much the founders own and whether they have pledged their shares against loans, but a proxy is not the thing itself. Likewise, it cannot tell that a whole industry is quietly being replaced. The accounts of a company in a dying business look excellent right up until they do not.
The business numbers are only as fresh as your last export
Prices update every trading day. The company financials update only when you download new files from screener.in and put them in the downloads folder. If you exported six months ago, the app is judging today's price against half-year-old accounts. Re-export at least every quarter, ideally after results season.
Every level is a probability, not a promise
Support is not a floor made of concrete. It is a price where buyers have shown up before and might show up again. Targets are not destinations the stock has agreed to visit. A target can be missed by a rupee and then never reached again. Treat all of them as reasonable places to have a plan, not as forecasts.
These scores are not backtested results
This matters and it is easy to miss. The scores look precise and rigorous, and they are calculated consistently, but nobody has run them over twenty years of history to prove they made money. The thresholds and weights are informed judgements, not values tuned against measured outcomes. A score of 82 does not mean that stocks scoring 82 have historically returned anything in particular. It means this stock ranks well on the measures the app happens to use.
Position sizing assumes you honour the stop
The "risk about 1 percent of your portfolio" arithmetic collapses the moment you do not sell at your stop. If you hold on hoping for a recovery, your 1 percent risk quietly becomes 3 percent, then 8 percent. The number was never a prediction of your loss. It was a description of a rule you have to follow.
Thin history and thin trading make everything less reliable
A recently listed company has too little history for the longer-term averages to mean much, and the app flags this and reduces its own conviction. A stock where few shares trade has unreliable support and resistance levels, because it takes very little money to move the price. Both cases are marked in the data quality note.
The app also looks for stocks stuck at an exchange price limit (a "circuit"): if the price was frozen at its limit on two or more of the last ten sessions, the verdict is Avoid, because there may be nobody to buy from or sell to. The exchange's own surveillance lists (ASM and GSM) are not part of the data this app reads, so a stock can be on one without the app knowing. Check the NSE website before buying anything small or sharply moving. During market hours the newest price is a live quote: volume is judged on completed days only, and a breakout needs the volume to show up, so most breakouts are confirmed after the close.
11. A worked example from start to finish
One imaginary stock, from noticing it to closing the position. The company is invented and so are all the numbers.
Monday: finding it
You have Rs 10,00,000 invested in total. You open the dashboard and click the Position Trader preset, because you are looking for something to hold for several months rather than several weeks. You drag the score slider's left handle up to 65 so the weak stuff disappears, and you tick the Volume Breakout screen.
Eleven companies remain. One of them, Bharat Widgets Limited, has a score of 78 and a green badge reading "Buy now". Its small price line climbs steadily from left to right with a step up at the very end. You click the name and the full chart opens.
Monday: checking it properly
Conviction is 76 and Confidence says High. Seven of the eight checklist items have passed. The one that failed is "Not stretched away from its average", showing the price is 9 percent above its 20-day average. That is under the 12 percent limit, so it is a near miss rather than a warning, but you note it.
The reasons list says the stock is trading 22 percent above its average price of the last 200 days, that it has beaten the market by 19 percent over the past year, that it broke to a new high on unusually heavy volume, and that it earns a 24 percent return on the capital it employs. There is no data quality warning, meaning the price history is complete and your export had the columns it needed.
You spend two minutes searching the company's name in a news site. Nothing alarming. You check the sector: industrials, and you do not already own an industrial. Good.
Monday: writing down the plan before touching the broker
The trade plan reads:
- Current price: Rs 500
- Entry zone: Rs 488 to Rs 505
- Stop loss: Rs 475, which is 5 percent below
- Target 1: Rs 550, which is 10 percent above
- Target 2: Rs 588, which is about 17.5 percent above
- Risk/reward at Target 1: 2.0
- Suggested position size: 10 percent of the portfolio
Ten percent of Rs 10,00,000 is Rs 1,00,000, which buys 200 shares at Rs 500. If the stop is hit you lose 200 times Rs 25, which is Rs 5,000, or half a percent of everything you have. You are comfortable with that number. This is the moment to decide, not later.
Tuesday: placing the order
The market opens and the stock is at Rs 507, just above your entry zone. You do not chase it. You place a limit order to buy 200 shares at Rs 502, which means you will only buy at Rs 502 or better.
By lunchtime it dips and your order fills at Rs 502. You immediately place a stop-loss sell order for all 200 shares at Rs 475, so that if the price falls there while you are asleep or at work, the sale happens without you.
You go back to the app and press "I Bought This", entering Rs 502. The stock moves into Active Positions.
The following weeks: managing it
You check once or twice a week, not hourly.
Week 1. The price drops to Rs 488. You are down about 3 percent. The Active Positions view says Hold, stop stays at Rs 475. Nothing has broken. You do nothing. This is the part that is psychologically hard and mechanically easy.
Week 3. Back up to Rs 528. Still Hold, with a note that the next target is 4 percent away. You are tempted to buy more. You do not, because that would push the position past the size you already decided was sensible.
Week 6. The price touches Rs 553, past Target 1. The Active Positions view changes to "Sell half, move stop to break-even". You sell 100 shares at Rs 552, banking Rs 5,000 of profit. Then you cancel your Rs 475 stop order and place a new one for the remaining 100 shares at Rs 502, your purchase price. From this moment the trade cannot lose you money.
How it ends, two ways
The good ending. Week 10 and the price reaches Rs 590. The view changes to "Sell (second target reached)". You sell the last 100 shares at Rs 589. Total profit: Rs 5,000 from the first half plus Rs 8,700 from the second, which is Rs 13,700 on Rs 1,00,400 invested, roughly 13.6 percent in ten weeks.
The ordinary ending. Week 8 and the market turns down. The price slides back to Rs 502 and your break-even stop sells the remaining 100 shares. You keep the Rs 5,000 from the first half and lose nothing on the second. A modest win from a trade that ultimately did not work.
The ending you must avoid. At week 6 you decide the stock is going much higher, so you do not sell half and you do not move the stop. At week 8 it falls to Rs 475 and your original stop sells everything for a Rs 5,400 loss. Same stock, same information, same app. The only difference is that you stopped following the plan you wrote down on Monday.
12. Common questions
Why is a stock I like marked "Avoid"?
Usually because one specific problem has overridden everything else. Open the stock and read the first line of the reasons list, which states the exact problem. The most common ones are that the price is below its average of the last 200 days, meaning the stock is in a downtrend regardless of how good the company is, or that too few shares trade each day.
Also remember what "Avoid" means: it is a statement about today, not about the company forever. A superb business in a twelve-month downtrend is still a superb business. The app is telling you this is not the moment to put money in, and quite often that turns out to be right.
Why is the score different from yesterday?
Three possible reasons. First, a new day of prices has arrived, so every average, every trend and every risk figure has been recalculated. Second, you may have moved the weight sliders or clicked a different preset, which changes the blend even though nothing about the stock changed. Third, you may have replaced a file in the downloads folder, changing the business figures.
Small daily movements of a few points are completely normal and mean nothing. A jump of twenty points overnight means something real happened, and is worth investigating.
What if the price gaps straight past my stop overnight?
It happens, and it is the one risk a stop loss cannot protect you from. Your stop sits at Rs 475. Bad news comes out after the market closes. The next morning the stock opens at Rs 430 and your order sells there, not at Rs 475. Instead of the Rs 5,000 you planned to lose, you lose about Rs 14,000.
You cannot prevent this. You can only make it survivable, and that is what position sizing is for. If any single stock is only 10 percent of your money, then even a terrible overnight gap is a bruise rather than a catastrophe. The people ruined by gaps are the ones who had half their savings in one company. Two further habits help: avoid opening a new position the day before the company announces its results, and spread your money across several unrelated sectors.
Why does it say a column is missing?
Because the spreadsheet you exported from screener.in did not include that piece of information. Screener lets you choose which columns to include, and the default selection does not cover everything this app can use.
The message is not an error and nothing is broken. That part of the business score is simply skipped rather than guessed, and the score is rescaled over what remains. The message lists the exact column names to add. Go back to screener.in, add them to your query, export again, and replace the file in the downloads folder. The more complete the export, the more reliable the business score.
Why is the first run so slow?
On the very first run there is no price history stored on this computer, so the app has to download the exchange's daily file for every past trading day it needs, unpack each one, and build a history for every stock. That is a large amount of downloading and processing, and it can take a while.
It only happens once. From then on, the app already has everything up to yesterday stored locally and only needs to fetch the one new day. Subsequent runs start in seconds. Let the first run finish rather than interrupting it.
Can I use this for intraday trading?
No, and you should not try. Every calculation in this app uses one price per day, the closing price, and everything is updated once a day after the market shuts. It has no idea what is happening at this minute.
The shortest horizon anything here is designed for is about one to four weeks, and that is the Momentum Swing preset. Using daily signals to make decisions inside a single session means acting on information that is already hours out of date.
What does it mean when there is no resistance level?
Resistance is a price the stock has repeatedly tried and failed to get above. If no resistance is shown, it usually means the stock is at the highest price it has ever reached, so there is no failed attempt above it to point at. Everyone who owns it is in profit, and nobody is waiting to sell just to get their money back.
The app treats this as a positive sign rather than as missing data, and scores it accordingly. It does have one practical consequence: with no ceiling in the way, the first target is set purely from the risk arithmetic, at twice the distance to your stop, rather than at a level the chart suggests.
Should I follow the verdict blindly?
No. The verdict is the summary of a calculation, not a judgement. It has never read the company's annual report, never heard of the lawsuit filed last week, and does not know that you already own three stocks in the same industry.
Use it as a filter, not as an instruction. A sensible routine is: let the app narrow thousands of companies down to a handful, read the checklist and the reasons for each, search for recent news, check the sector against what you already own, and only then decide. If you would not be able to explain in one sentence why you own something, do not buy it.
The table is empty. What did I do wrong?
Almost certainly you have ticked several screens at once. Remember that screens combine by narrowing: a stock has to pass every ticked chip. Press Clear all and add the filters back one at a time. Also check the Filters and Query drawers, since a badge on either button means something is still switched on in there, and that the sliders have not been left at a narrow range.
How do I know I am looking at today's prices?
The chip in the top right corner always says which trading day the numbers
come from, for example Prices: Monday 17 August. Green means it is the most recent close
that exists. Yellow means it is waiting for this evening's file. Red means the data is genuinely
old, and a banner across the top will explain why.
You do not have to do anything to update it. The exchange publishes one file per trading day
after the close, usually between six and half past six in the evening, and the app watches for
it and recalculates everything the moment it appears. If the file is late it keeps checking
until nearly midnight.
Two normal situations look like being behind but are not. At the weekend the newest close is
Friday's, because that is the newest there is. And on the day after a public holiday the app
will briefly say it is one day behind, because no file is published on a holiday and there is no
reliable free list of Indian market holidays to check against. In both cases it sorts itself out
on the next trading evening.
How often should I look at this?
Once a day at most, after the market has closed, since that is when new data arrives. For most people once or twice a week is plenty. Nothing in this app changes during trading hours, so refreshing it at eleven in the morning shows you exactly what it showed you at nine. Checking constantly does not produce better decisions, only more of them.
13. The Market mood tab
Every other part of this app asks "is this a good company at a good price". This tab asks a completely different question: is now a sensible time to be buying anything at all?
The two questions are separate, and mixing them up is one of the most expensive habits in investing. A genuinely excellent company bought three weeks before a market-wide fall still loses you money for a year. Knowing what sort of market you are in will not stop that happening, but it will change how much you put in.
The big number
One score from -100 to +100, and one of four words:
| Risk-on | The market is healthy. A normal environment for the Buy verdicts in this app. |
|---|---|
| Neutral | Mixed. Nothing argues for being either aggressive or defensive. Be more selective and take your time. |
| Caution | Weakening. Smaller positions than suggested, more cash than usual, stop losses treated as firm. |
| Risk-off | Real stress. This is where most individual investors lose the most, usually by averaging down. Buying can wait. |
Where the number comes from
Roughly four fifths of it is measured from Indian market prices, and every input is listed underneath with its own weight and its own explanation. Nothing is hidden inside the total. The inputs are:
- Where the index sits against its 200-day average, and whether that average is rising. The oldest definition of a market's direction there is, and its great merit is that it is not a matter of opinion.
- Whether the 50-day average is above the 200-day. Whether the recent trend agrees with the long one.
- How far the market is below its 52-week high. Note that past about 25% down this stops getting worse rather than continuing to fall. Deep declines have historically been followed by better returns, not worse ones, which is the opposite of how they feel at the time.
- How violent the market has become, measured against its own history rather than a fixed number, because what counts as a wild day changes over the years.
- Its direction over three months, with the last week excluded so one loud day does not flip the reading.
- Breadth: how many individual shares are above their own 200-day average. This is the most useful number on the page. An index can reach a new high carried by a handful of enormous companies while most shares are already falling. Breadth is what shows you that.
The world outside India
About another sixth comes from prices in markets that move Indian shares whether you follow them or not: American equities, the US ten-year yield, the rupee, oil, gold and Hong Kong. Each is shown with what a move in it actually means here. India imports most of its oil, so a sustained rise in crude feeds straight into inflation and company margins. A weakening rupee makes Indian assets worth less to a foreign investor, who often sells.
The headlines, and why they count for so little
The tab also counts hopeful and fearful words across recent market and geopolitical headlines. This is included because how the world is being reported does carry some information. It is capped at ten points of the score, and labelled as weak wherever it appears, because counting words in a headline is not the same as understanding what happened. A number like that is exactly the sort that sounds authoritative and is not, so it is kept on a short leash. There is a test that proves the cap holds even when the headlines are as hysterical as they can possibly be.
What to do with it
Use it to size positions, not to time the market. Selling everything in a downturn and buying back later is, for almost everyone, worse than doing nothing at all: you have to be right twice, and the best days often arrive in the middle of the worst months. What this tab is genuinely good for is stopping you from putting your usual amount into a high-beta stock in a week when the market is falling apart.
When you open a stock you will also see a line about market conditions next to its verdict, using that particular stock's beta. A beta of 1.4 means it has historically moved about 40% more than the index in both directions, and in a falling market only one of those directions matters. That line does not change the stock's score. A number that silently moves with the weather could not be explained to you, and every score in this app is meant to be explainable.
It runs by itself each evening, once that day's closing prices have actually arrived rather than at a fixed time, so it is never computed on yesterday's numbers. You are always reading a saved result, which is why the tab opens instantly.
14. The "Test one stock" tab
This tab answers a question every tool like this one avoids, about a company you actually care about: if you had followed this app on THIS stock, would you have done better than simply buying it and never touching it again?
It is allowed to say no. It has been written so that it can.
What it does
You give it a symbol and a period. It goes back to the start of that period and, at each review date, works out what this dashboard would have told you that morning — using only the data that existed on that day. Then it simulates three ways of owning the stock, on identical terms, paying identical costs.
| Buy and hold this stock | Buy on the first day and never sell. This is the bar. It takes one purchase and no further thought, so anything more complicated has to beat it to have been worth doing. |
|---|---|
| Follow the verdict | Hold the stock whenever the dashboard rated it a buy on the review date, and sit in cash the rest of the time. |
| Follow the full trade plan | Buy on a Buy verdict, then leave on whichever comes first: the stop loss, the first target, or a downgrade. The stop and target are checked against every day's high and low, not just review dates, because that is the only version you could actually have followed. |
| Machine learning model | Optional, and slow, because the model has to be trained across the whole market first. See below for why it is not trained on your stock. |
Why the model is not trained on the stock you asked about
It is the obvious thing to do and it is the wrong thing. Three years of daily bars is about 750 rows, and once you remove the overlapping ones it is nearer 35 genuinely independent observations against twenty-odd inputs. A model fitted to that memorises the particular sequence of those years. It would score brilliantly and mean nothing, because the thing it has overfitted to is exactly the history you are testing it on. So it is trained across the whole market and merely applied to your stock.
One consequence worth knowing: the model is trained to pick stocks that beat the median stock, not to decide whether to be in the market at all. In a month when everything fell, beating the median still means losing money.
Why most backtests you see are worthless
It is very easy to build one that promises 40% a year and delivers nothing. These mistakes account for almost all of it, and this one is built specifically to avoid each:
- Knowing the future. The single most common error, and the only one that makes a backtest look better rather than broken, which is why nobody catches it. Here every verdict is recomputed from history cut off at that date, using the same code the live dashboard uses. There is a test that appends an invented future to the data and checks that not one verdict in the past changes.
- Ignoring costs. Securities transaction tax, exchange fees, GST, stamp duty, the flat depository charge on every sale, and an allowance for slippage are charged on both sides of every switch. A strategy that trades often can look profitable before costs and be a loss after them.
- Using today's financials. Your Screener export holds current figures, so using it to judge the stock in 2023 would mean the test already knew how the company turned out. The business quality score is therefore switched off for this test.
- Assuming stop losses fill at the stop. They do not when a stock gaps through overnight, and gaps cluster around bad news. Here a gap fills at the opening price, and when a single day's range covers both the stop and the target, the stop is assumed — daily data cannot say which came first, and guessing the other way invents profits.
One more, in the other direction: money not invested earns interest. Cash is credited at 6.5% a year, on the basis that it would sit in a liquid fund. Without that, a strategy selective enough to sometimes hold nothing would be punished for exactly the behaviour you want from it.
The two things the answer will not tell you
It is a sample of about five. One stock over a few years is a handful of decisions. Whichever way it comes out, one lucky exit would usually have reversed it. The page says how many completed trades the result rests on, right next to the result, for that reason.
You already know how the story ended. You picked this stock today, knowing what happened to it. That is not a flaw in the test, it is a property of choosing what to test backwards, and no amount of careful simulation removes it.
Tax, which is not in any of the numbers
Buying and holding defers capital gains tax indefinitely. A strategy that trades realises it every time it sells. That asymmetry is real, it always favours holding, and it is sometimes decisive. Rather than bury it, the page counts the gains each strategy actually realised and tells you what the tax on them would have been, so you can subtract it yourself.
If it says the price history looks wrong
Believe it, and do not read the numbers underneath. When a company does a 1-for-1 bonus, the share price halves overnight and nobody loses anything — you simply own twice as many shares. The older prices have to be rescaled to match, and for a sizeable minority of stocks in the stored history that has not happened. The result looks exactly like a 50% crash.
That matters here more than anywhere else in the app, and it matters in one direction: holding through the imaginary crash looks catastrophic, while any strategy that happened to be in cash that week looks brilliant. So the page refuses to give you a headline answer when it spots one, and shows you the day and the ratio instead.
A closing reminder
This is a personal tool for organising your own thinking. It is not investment advice, it is not regulated, and it comes with no guarantee that any number in it is correct. You are responsible for every order you place. Invest only money you can afford to lose, and when something here disagrees with your own good judgement, trust your judgement.
What sort of market is this?
Test one stock
What to test
Dealing costs are charged in full on both sides of every switch, and money sitting out of the market earns a liquid fund rate rather than nothing. Tax is not modelled, and that always favours holding: the summary says by roughly how much.
Watch
Alerts
Reject
Company news
Pick a stock from the table and press News to see recent headlines about that company, from Google News. Where a headline or snippet states a plain fact - a result, a deal, a number - it's pulled out and highlighted below the headline. Analyst opinions, price targets and predictions are left out of that list on purpose; this is a simple keyword filter, not a fact-checker, so treat it as a starting point and use the link to read the full article.
Vedic financial astrology
Today's sky
Good stocks, right now, in a favoured sector
This does not use astrology to judge whether a stock is worth buying - that is the dashboard's job, same as everywhere else in this app. It only adds timing: among the stocks already rated Buy on dips or better on their own fundamentals and technicals, which sit in a sector currently favoured by a Major (multi-year) planetary transit, not a passing one.
Coming transits
Walked forward from today. Major, multi-year transits are shown by default; shorter-lived minor ones are a click away.
Ten-year backtest
Samples the sky roughly once a month over the available history, scores every sector, and checks that against what each sector actually returned over the following month. Reports the correlation honestly, including when it finds none.