01 / The problem

Most investing is guesswork.

Chasing trends. Reacting to news. Trusting a good story.

Without a way to test an idea against real Indian market history, even a careful investor is only guessing.

There is a better way

02 / The standard

There is a right way.

Own good businesses. Hold them for a long time.

That is how lasting wealth has been built. Wealth comes from businesses compounding, not prices moving.

the business the price

03 / The question

But how do you know it works?

You can actually test it. Run it through years of real history. See exactly what it would have made, year by year.

04 / The wall

And that is brutally hard.

Gather data on thousands of Indian companies, decades deep. Clean it. Rebuild every measure, every company, every year. Construct the portfolios history would have built. Track them. Then test the same idea from every starting point you can find, so one lucky stretch does not fool you.

Did it actually work?

Almost no one can answer that honestly. Not alone.

05 / Krest

The right way, finally made practical.

Turn an investing idea into a fully tested strategy, without months of data gathering, spreadsheets, or coding.

Krest ›Directory ›New strategy Cancel ✓ Validate
Composer ready 0 lines

The hard part is having the idea. Krest handles the rest.

See how easy it is to build a strategy
See what it finds

Or skip straight to the answers.

Krest gives you access to a growing library of strategies, each tested against real history.

The signature insight

The first result is rarely the whole story.

A strategy that looks incredible from one starting date can look very different from another.

Rolling Performance · CAGR by start date · Quality with low debt
Horizon 1Y 3Y 5Y

Example analysis Quality with low debt. Ten years of history today.

Median 5Y return
21.2%
a year, across 53 starting points
Best start
38.0%
a year, if you had started at the low
Worst start
11.4%
a year, if you had started at the high

That’s why Krest shows the entire range of outcomes.

06 / The depth

Then it shows you more than you expected.

How reliable it was. How painful it got. What it actually owned. Where it worked. And where it didn't.

The one that matters most
06 · b / pain

Where did it hurt?

The worst stretch the idea ever put you through, and how long it took to come back. Seeing this before you commit is what lets you actually hold on when it happens.

-21.4%
worst stretch
14 mo
to come back

Drawdown vs benchmark

The headline view growth of ₹100 · all variants · Jun '16 onward
Variant All
Return & growth
14.2%
a year, median
Reliability
1.18
risk-adjusted
Worst stretch
-21.4%
peak to trough
Sorting strength
0.34
weakest to strongest
Names held
22
rebalanced quarterly
D10 (signal) 5.7x Top 30 5.4x Top 10 3.4x Benchmark 3.6x

Example analysis The headline view. Quality with low debt. Ten years of history today.

The same answer, seen through many lenses

06 · a / consistency

Did it work consistently?

Every five-year window in history. How risk-adjusted return held up.

Rolling 5y Sharpe

06 · c / sorting

Does the idea sort cleanly?

Weakest to strongest on the idea. How each group did.

Decile staircase

06 · d / size mix

What size of companies?

Large, mid, small. How the mix shifted.

Size composition over time

06 · e / sector mix

What sectors?

Where it leaned, year by year.

Sector composition over time

06 · f / regimes

Through good times and bad?

Across growth, inflation, interest-rate cycles. How it held in each.

Performance by macro regime

06b / Across strategies

And it does not stop at one strategy.

Compare any idea against any other. Through the same lenses.

Risk and Return return above the market · against how bumpy the ride was
View Risk vs Return
Up and to the left is better: more return above the market, with a calmer ride. Each cloud is one idea tested from many starting points; the bright dot is its middle.

Example analysis Three ideas. Ten years of history today.

Quality with low debt 14.2% a year Steadiest Financially strengthening 17.6% a year Fastest Quality at a fair price 12.8% a year Safest
behind the login

And many more views, every strategy, every lens, behind the login.

07 / The category

Not a screener. Not a trading tool.

A screener
only shows what looks good today.
A price tool
only shows what the price has done.
Krest
shows what actually worked, across years of real Indian market history.

Why these results deserve your trust.

25 years · 5,000+ companies · every price, every dividend, every set of accounts

Your free trial runs on the most recent 10 years.

Here’s what we did to keep them honest.

We test the market as it was, not as it looks today.

Companies disappear all the time. They get bought, they merge, or they go under. A test built from only the companies that still exist today has quietly kept the survivors and dropped the rest.

why this matters ↻

Keeping only the survivors flatters the result.

The companies that vanished were real investments at the time. Leaving them out makes the past look safer and richer than it really was. Krest includes every company that was trading on each date, even the ones that no longer exist.

survivorship-bias-free registry of NSE and BSE equities, including long-delisted names · 5,000+ companies, 1,000+ no longer trading, all retained · a name delisted on one exchange is still followed on the other · point-in-time universe rebuilt for each date · every name eligible only on the days it actually traded

A number can be exactly right and still mean nothing.

Financial data is messy. A missing figure that ranks as a winner. A loss-making company that sorts to the cheapest in the market. A tiny base that turns a small gain into “three thousand percent a year.” Each one is arithmetically correct. Each one is meaningless.

why this matters ↻

We remove these before they can rank.

A ratio with a negative or near-zero denominator. A growth rate off a base close to nothing. A blank cell read as a zero. Any of these can reach the top of a ranking and quietly ruin a strategy. Krest spots them and leaves them out, so a company is scored only when the number truly means something.

negative and near-zero denominators disqualified · growth rates and ratios off a very small base guarded, not ranked as real · missing data never becomes zero, the company is simply left unscored on that measure · 100+ base measures, each traceable to a line in the company’s own accounts · every published figure independently re-derived from source and reconciled

We only use what was known at the time.

Every decision is based only on the information available on that date, nothing that came later. A test can quietly use information that did not exist yet: a result before it was public, or a value known only later.

why this matters ↻

Hindsight makes any strategy look brilliant.

So we never allow it. A company’s results are used only after they were actually announced. Each choice is judged only on what an investor could genuinely have known at the time.

no look-ahead · company results used only after their announcement date (fiscal year end + 60 days, the filing deadline) · declared limit: a later restatement is not separated from the original filing

A price chart is not the same as your return.

A split or a bonus can make a normal day look like a crash. Years of dividends never show up on the chart at all. What you actually earned is a different number.

why this matters ↻

We rebuild the return you would really have received.

Krest adjusts every price for splits and bonuses and adds back every dividend, so the figure is the actual money you would have taken home. It is then compared against a benchmark that also counts its dividends. A fair comparison.

total-return series: splits and bonuses applied on the ex-date, dividends reinvested, NSE and BSE reconciled · benchmark: NIFTY 500 Total Return Index, dividends counted on both sides · demergers left unadjusted (~50 cases), which works against our own numbers, not for them

There is always one start date that looks amazing.

Pick the right month to begin and almost any strategy looks great. Show only that single run, and the number means nothing.

why this matters ↻

So we test every start date, and every holding period.

Krest runs the strategy from every starting month across the ten years, and from each start it measures how you would have done over one, three, and five years. You see the full range, the worst outcome beside the best. If a strategy only works from one lucky month, or over one convenient period, you see it here.

evaluated from every reconstruction month in the decade, none dropped · each start measured over 1, 3 and 5 year holding periods · reported as the median with the full P10 to P90 range, never a single run · worst outcome shown beside the best

One thing we’d rather be upfront about.

Every return on this page is before trading costs and tax, so the figure you keep is lower. For a strategy that replaces most of its holdings each year, that gap is roughly two to four percent a year.

We will never tell you what happens next.

Everything above already happened. None of it is a promise. Anyone who promises you the future is the one to trust least. That is why you see a range of outcomes, not one confident number. An honest test says what a strategy could have done, then refuses to pretend it knows what it will do.

Prices, splits and dividends from the NSE and the BSE. Company accounts from their own published annual reports. Delisted and suspended names from the exchanges’ own records. How we handle the data →

08 / The way in

The way investing was meant to be done.

See whether your ideas survive India's market history. One free account. One door in.

Step inside

Free during beta. No card. 25 years of data behind it — your trial runs the most recent 10.