Know what a strategy is actually worth before you risk anything on it.
QuantSmart Technologies develops institutional-grade quantitative strategy testing and empirical simulation architecture for Indian equities. Built on this foundation, SmartEdge 3.0 is our debut first product — a zero-bias backtesting engine engineered to compute what any systematic rule would have genuinely retained over a decade of Indian market history.
Define any rule — how stocks are ranked, from which universe, how capital is distributed, and how often it is rebalanced. SmartEdge 3.0 calculates what you would have kept after every statutory charge, exchange fee, slippage hurdle, and rupee of Indian capital gains tax.
Nothing is pre-selected. The rule is yours; the arithmetic is ours.
- Slippage, STT, stamp duty, exchange and SEBI fees, GST, DP charges
- Capital gains tax, netted across each financial year
- Loss set-off and carry-forward under Indian rules
- Whole shares only, with undeployed cash carried forward
- A liquidity floor, so results stay tradeable
- The benchmark taxed on its own correct basis
SmartEdge 3.0 — Institutional Investment Analytics
SmartEdge 3.0 is QuantSmart's debut flagship product. Purpose-engineered to dismantle flattering retail backtesting assumptions, survivorship illusions, and unmodeled friction, it delivers rigorous, empirical verification of systematic equity strategies across Indian equities with absolute mathematical fidelity.
Dual-Factor & Valuation Scoring
Score assets with empirical Momentum (Rate of Change, Volatility-Adjusted return, Sharpe-like daily ratio, Log Regression Slope with R² trend quality) alongside deep Fundamental Valuation metrics (P/E, P/B, EV/EBITDA, Earnings Yield, and Dividend Yield).
Adaptive Regime & Trend Gating
Safeguard capital during severe drawdowns with automated moving-average regime filters (50-day EMA, 200-day EMA, and Golden Cross gating), paired with custom RSI boundary limits to reject overbought or decaying stocks.
Survivorship-Clean Universes
Seamless testing across Nifty 50, Midcap 100, Smallcap 250 (turnover-derived to avoid survivorship bias), and Sector Rotation / Composite models across 2,400+ NSE equities over ten years of point-in-time data.
Statutory Friction & Indian Tax Engine
Every rupee accounted for: STT, Stamp Duty, Exchange/SEBI fees, GST, and DP charges, plus annual Indian Capital Gains Tax (STCG @ 20%, LTCG @ 12.5%) with financial-year loss set-off and carry-forward rules.
Discrete Execution & Cash Tracking
Eliminates theoretical decimal shares through exact integer share allocation, enforces liquidity volume floors, factors in empirical bid-ask slippage, and accurately compounds unallocated cash balances period by period.
Audit Ledgers & Benchmark Analysis
Complete institutional transparency with granular trade-by-trade ledgers, peak-to-trough drawdown profiling, periodic rebalance scorecards, and direct performance comparisons against correctly taxed benchmark indices.
Why the number matters more than the strategy
The same momentum rule, run twice on the same decade of prices. The first run allowed it to buy anything. The second required that somebody was actually selling in the size being bought.
Both runs are correctly calculated. The first quietly assumes a willing seller at the closing price for any quantity, which is not how a market works. The second is the number you could have kept.
And the charges nobody models
Most published backtests report gross returns. The charges below are paid on every position at every rebalance, and capital gains tax is deducted on top of them.
Across our own testing, tax alone consumed roughly 60% of the amount by which a quarterly strategy beat its benchmark.
What every result deducts
Below is a single ₹20,000 position, bought and sold at exactly the same price. Nothing moved, and this is what came back.
The price did not move. Rebalancing quarterly, this is paid four times a year on every position; monthly, twelve.
Three tools
Each answers a different question.
1st Product
SmartEdge 3.0 — Flagship Engine
QuantSmart's debut first product. Apply any quantitative rule to a decade of Indian equity prices and see what it would have returned after all slippage, statutory charges, and capital gains tax, against real taxed benchmarks.
Launch SmartEdge 3.0Scores
The full ranking your formula produces at every rebalance date, so you can set weights by rank and see which companies filled each slot period by period.
Open scoresScreener
The same formulas and filters applied to the latest data instead of a historical window. Available to registered users.
Sign inSome of what the testing found
These are results from our own work, including the ones we would rather not have found. They are the reason the tool is built the way it is.
Plans
The free plan keeps the full cost and capital gains tax model — that is the finding, not a feature. What the paid plans add is breadth.
Free
Nifty 50, rate of change, quarterly, five equally weighted positions. Five backtests a day.
- Every charge and tax modelled
- Benchmark taxed on its own basis
- Trade ledger and drawdowns
Pro Most people want this
The whole platform — every universe, formula and frequency, plus the screener and scores view.
- Midcap, smallcap and sector universes
- All momentum formulas and value metrics
- Custom weights, entry filters, SIP
Research work — priced per engagement
Significance testing, walk-forward validation and Monte Carlo against random selection are analyses we run and report on rather than features of the website.
What we do not do
We do not tell you what to buy. We do not publish tips, model portfolios or target prices, and we do not sell a strategy. QuantSmart is not registered with SEBI and does not give investment advice.
What the platform produces is arithmetic on a rule you defined. A score describes what a price has already done. It is not a forecast, and the testing above is the reason we are careful about the difference.
Before you rely on any result, including ours. Look at what the rule did in its worst period rather than its best. Check how often it beat the benchmark, not by how much on average. And ask what result would have made you abandon it — if there isn't one, the test was not a test.