Glossary

Measures how one-sided recent moves have been. Overbought means strong, not doomed.

Deliberation should be proportionate to this, not to the amount involved. A ₹15 lakh index fund purchase can be undone on Tuesday; ₹3 lakh of booking money on an under-construction flat cannot be undone at any price you would accept.

The fastest Ichimoku line. Crossing it means little on its own; where price sits relative to the cloud decides everything.

The single most useful cross-check on an income statement, and it needs two numbers you already have open.

A lengthening cycle is often the first quantitative sign that business quality is slipping.

Most of the benefit is captured by about fifteen genuinely uncorrelated positions.

It is why mean reversion has a genuine basis on an index and a shaky one on a single stock.

Peter Lynch’s word for a textile company buying a hotel chain. Competence transfers across industries far less than boards assume.

Why "I will sell when it returns to my buy price" is such a costly sentence.

The rubber band. Wins often, loses large, and works only where the fall was movement rather than information.

Embedded value assumes policies run their full term. Strong VNB growth alongside falling persistency is value being reported that will never be collected.

Each regret produces the opposite error next time, which is how it compounds.

True of the annualised return, which converges roughly with the square root of the horizon, and false of the final amount, whose spread widens over the same years. Most arguments about it are two people each defending one half.

Some names trend for weeks and some reverse in days — reliably, over years.

Not a sale but a recovery process, run on notices rather than on screens. The chart usually stopped months earlier, when the security was suspended.

The test is territorial, not visual: if the pause stays inside the ground the trend already won and volume thins while it happens, it is a rest. If it takes that ground back on rising volume, the name of the shape stops mattering.

Same fund, same manager, same portfolio — typically 0.5–1% cheaper every single year.

Cumulative operating cash flow divided by cumulative profit over five years is the quick version. Above 0.8 is healthy.

The failure is often a stronger signal than the pattern would have been.

The part of a high multiple carrying most of its value. A company can beat next year’s estimate and still de-rate, because the beat answers the rate while something has changed the market’s view of the duration.

The most misunderstood word in trading. In a strong trend RSI can stay above 70 for months.

It describes momentum, not value. In a strong downtrend an oscillator can print oversold all the way down, and bounces top out near 55–60 rather than 70.

In broking, prefer the narrower DDPI, which permits debits only for settlement, over a broad POA. In family finance, know the limit: Indian agency law treats an agent’s authority as ending if the person who granted it becomes of unsound mind, so an ordinary POA is generally understood not to survive the loss of mental capacity — the very case families buy one for.

The bet is on duration rather than cheapness. Its failure mode is overpaying — a superb business bought at an extreme multiple can be dead money for a decade while earnings catch up.

In a strong uptrend pullbacks bottom near 40–50, not 30. Waiting for the textbook oversold reading in a trending stock means never buying at all.

Same scheme, same manager, same portfolio as the direct plan, typically 0.5–1% dearer every year. The extra is charged whether or not any advice is ever given.

Shows up as excess cash and an inability to deploy a lump sum. It can also flip into compulsive spending.

If you cannot build the case against your own position, you do not understand it well enough to hold it or to leave it.

The horizon that was ten years becomes ten days. It is why stopping for the day after a significant loss is protection rather than punishment.

Beta measures your exposure to it. A high-beta portfolio carries it without borrowing.

The textbook version has the agent taking too much risk. Inside a family it usually inverts: the person managing a parent’s money carries the blame without the loss, and takes far too little.

Usually described as retail being front-run by machines. The accurate version is narrower: algorithms compete for very short-term moves, which makes intraday harder and barely touches someone holding for months.

Only meaningful if it reflects what you would otherwise have done. Use total return versions.

Reducing exposure in a bubble means underperforming visibly for a long time. There is no version that avoids that.

The most forecastable dilution there is: the number of shares that will exist on conversion is public from the day the general meeting approves it.

Diversification helps least exactly when it is needed most, because in a panic people sell what they can rather than what they want to.

Only useful if the objections are allowed to stand. A token version leaves you more confident, not better calibrated.

The version to use, because headlines quote the other one. A wide gap to basic EPS means a large claim on your ownership exists and has simply not been triggered yet.

A warning to tighten risk, not a trigger to reverse. It can persist for weeks.

Conviction, hesitation, handover. Look at almost any stock that topped out badly and some version of this shape is there.

Real, persistent, and it needs a catalyst to close. Being right without one pays nothing.

Lower volatility, milder gaps, stronger mean reversion, almost no company-specific risk. Different enough to justify genuinely different rules.

RSI, Stochastic, Williams %R, CCI and the MACD histogram all agreeing is one opinion reported five times. New information has to come from a different input — volume, breadth, relative strength.

Negative is the normal state for a company still building something. Persistently positive usually means assets are being sold, which flatters this year and shrinks the next one.

The climber's turnaround time. Written on a calm Sunday so the version of you reading panicked commentary has an instruction rather than a decision.

Persistent across markets and decades, with no comfortable explanation. It is a statistical tilt, not a prediction about any one stock.

RSI with the money attached. Where RSI asks whether it went up, this asks whether anyone paid for it to.

Forced diversification across sizes. The manager is legally unable to retreat into largecaps during a smallcap crash — which is the whole difference from a flexicap.

The honest version cuts both ways: a thesis you can barely defend gets a size that can go to zero without mattering.

Trend systems want expanding volatility; mean-reversion systems want it settled. Neither is broken when the regime changes — it is just out of season.

It manufactures revenue that never becomes cash. Where it surfaces is the related-party note and large receivables from group companies that persist year after year.

Put both after tax and the gap is usually far narrower than the headline version suggests. Which tax regime you are on moves the hurdle by points, not decimals.

The honest version of ROE. It cannot be manufactured with leverage.

The terms are published, the timeline is fixed and the outcome is largely arithmetic. They persist because they are boring, small and time-limited, which keeps large funds away.

It asks a different question from RSI — where you finished, not how one-sided the moves were. In a trend it pins at an extreme for weeks, so it belongs in ranges only.

Diversification is defined by whether drawdowns coincide, not by whether the rules look different.

What mean reversion trades away its high win rate for. The one position that never comes back is the whole risk in that style.

Where trend systems earn their money and mean-reversion systems get run over.