COST-FIRST ANALYTICS · NSE INTRADAY

Every trade starts below zero.

ThreeTwenty is cost-first analytics for Indian intraday trading. Every signal is priced net of brokerage, STT, GST, stamp duty and modelled slippage — before it reaches you.

0.146% round-trip friction at retail size, all charges and modelled slippage in
Join the waitlist No spam, no tips, no promises.

Named for 3:20 PM, the minute Indian brokers force every intraday position closed. The system is named after its deadline.

09:00
PRE-OPEN

Priced before the bell.

Most intraday traders in India lose money. The tools they are sold discuss entries, indicators and targets — and stay silent on the one number that is certain before any trade begins: what the trade costs. Nobody shows the trader their contract note first.

~7 in 10 individual equity intraday traders lost money in FY23, per SEBI's study.
~9 in 10 individual F&O traders lost money, per SEBI's study.

ThreeTwenty starts where the loss actually happens: the charges. A signal is not allowed to exist until it has paid, on paper, every rupee of brokerage, STT, exchange transaction charges, SEBI and IPFT fees, GST, stamp duty and modelled slippage.

09:15
THE OPEN

The arithmetic

A round trip at retail size costs about 0.146% of notional — 0.1063% in charges alone, the rest modelled slippage. Brokerage stops scaling at roughly ₹66,667 per side; at any size the friction floor is about 0.076%.

Your opponent isn't the market. It's your contract note.
THE LADDER — FRICTION VS. STOP WIDTH
STOP WIDTH
FRICTION AS % OF R
WIN RATE TO BREAK EVEN AT 2R
0.107 %
137 %
79 %
0.125 %
117 %
72 %
0.208 %
70 %
57 %
0.585 %
25 %
42 %
0.875 %
17 %
39 %
2.05 %
7 %
36 %

With zero costs, a 2R trade breaks even at a 33.3% win rate. Tight stops don't reduce risk — they hand it to the exchange. A 0.1% stop pays 137% of its own risk in charges. No indicator survives that.

True-cost calculator

NSE EQUITY INTRADAY · MIS

Rates follow a discount-broker equity intraday schedule. Charges are illustrative; verify with your broker.

Stop losses don't reduce risk. Tight ones donate it.

COST AS % OF R25%
BREAKEVEN WIN RATE41.7%
ZERO-COST BREAKEVEN33.3%
VERDICTPASSES

Cost ÷ stop width = the share of your risk unit the exchange keeps. Above 50% the system calls the trade uneconomical and it is never signalled.

2002
OAKLAND

The Moneyball math

A team with no money beat teams with all of it by pricing what the market mispriced. Indian intraday has the same mispricing: everyone models the signal, nobody models the toll. The edge, in the design document's words: "it is arithmetic, not signal quality."

THE PAYROLL CONSTRAINTTHE FRICTION MODEL

0.146% round trip at retail size, brokerage cap at ≈₹66,667 per side, 0.076% floor at any size. The budget nobody else prices.

ON-BASE PERCENTAGECOST AS A FRACTION OF R

One division: friction ÷ stop width. A 0.1% stop pays 137% of its own risk in charges. The unglamorous number that decides everything.

WINS ABOVE REPLACEMENTEXPECTANCY IN R, NET

Every strategy is measured against the free alternative: not trading. If it cannot beat zero after charges, it is replaced by zero.

SCOUT BIASWHAT WE SCREEN INSTEAD

Watchlist admission is by measurable criteria. Stocks that fail are excluded, not argued about.

SMALL-SAMPLE HUMILITYSTATISTICS THAT PUNISH US

Strategies are ranked by their lower confidence bound (z = 1.645) — the worst plausible version of themselves, not the best day.

TEAM BUILDINGTHE ROSTER

Book exposure split 55 : 35 : 10 across volatility buckets. Daily quota 2 / 2 / 1 per bucket, never backfilled.

DEFENCE WINS CHAMPIONSHIPSVETOES AND CIRCUIT BREAKERS

Six vetoes between trigger and signal, economics included. The daily-loss lockout survives a restart.

THE MINOR LEAGUESTHE PROVING GROUND

Every strategy proves itself against live prices, charged full modelled friction, before promotion. By statistics, not vibes.

IN SESSION

Built to say no.

Candidates enter a nine-stage funnel. Admission to the watchlist is by measurable criteria — liquidity, spread, tradable range — stocks qualify, they are not picked by vibe. Between "triggered" and "veto survived", six vetoes run. Economics is one of them: a trade that cannot pay its own charges is rejected the way a broker's risk system rejects an order.

Some days the right number of trades is zero.
55 : 35 : 10book exposure across volatility buckets
2 / 2 / 1daily pick quota per bucket, never backfilled
6vetoes, economics included
THE
VETOES

Watch a trade earn its signal.

Between a trigger and a signal stand six vetoes. Every candidate is walked through them live — and the cost of the trade decides the verdict before conviction gets a vote.

All six vetoes survived. Signal issued, priced net of charges.
1LISTINGadmission criteria met
2RMSexposure within 55 : 35 : 10
3BASKETquota 2 / 2 / 1 available
4CIRCUITdaily-loss lockout clear
5CLOCKrunway before 15:20
6CONTRACT NOTEcost ≤ 50% of R
COST AS % OF R 25%
SIGNAL ISSUED
>100% — EXCEEDS RISK >50% — UNECONOMICAL <50% — CAN PAY ITS COSTS

A trade whose friction exceeds 50% of its risk unit is never signalled.

Illustrative sequence. Costs computed by the friction model; the second pass shows a candidate rejected by the economics veto.

15:20
SQUARE-OFF

The deadline that names the system.

At 3:20 PM, Indian brokers force every MIS position closed. That constraint defines intraday trading in India — so the system is built around it, and named after it. Eight modules, each named after the thing it answers to.

CONTRACT NOTEThe charge model; predicts your broker's bill to the paisa.
RMSThe veto stack; rejects trades the way a broker's risk system rejects orders.
CIRCUITDaily loss lockout that survives a restart.
LISTINGWatchlist admission criteria; stocks qualify, they don't get picked by vibe.
BASKETThe daily picks, quota-capped per volatility bucket.
SETTLEMENTThe nightly recalibration after close.
PROVINGThe proving ground; strategies earn promotion against live prices, all charges in.
AUDITThe evidence layer; limited discoveries, corrected for false positives.
15:30
AFTER CLOSE

The evidence desk works nights.

RANKED BY THE WORST CASEStrategies are ranked by lower confidence bound (z = 1.645), not by their average or their best run.
CORRELATED DATA COUNTED HONESTLYOverlapping trades are not independent evidence. Effective sample size is computed and reported.
A FIXED BUDGET OF TESTSStatistical tests are budgeted in advance with false-discovery-rate correction. You cannot torture the data if the data has a lawyer.
OUT-OF-SAMPLE IS SACREDOut-of-sample results are printed but never used to pick winners.
Most backtests are written by the marketing department. Ours is written by the defence.

Every strategy proves itself against live prices, every fill charged the full modelled friction, before it is trusted with capital. Promotion is by statistics, not vibes — and until the statistics exist, the site claims nothing.

FAQ
Do you guarantee returns?
No. Promising fixed or certain returns is illegal in India's securities markets. We publish our method, our cost model, and our statistical rules. That is the whole offer.
Is there a track record?
Not yet. Strategies are currently proving themselves against live prices, net of all charges. When results exist, they will be published with confidence bounds and sample sizes.
Why so few signals?
Most candidates cannot pay their own costs. The daily quota is 2 / 2 / 1 across volatility buckets, never backfilled. Some days the right number is zero.
Why do costs matter at my size?
Below roughly ₹66,000 per side, charges are a fixed percentage of the trade. The smaller the trade, the larger the toll.
Is this investment advice?
No. ThreeTwenty is an analytics tool. It prices and filters candidate trades; it does not advise.
Which broker and segment?
NSE equity intraday (MIS), priced on a discount-broker charge schedule.
WAITLIST

When the statistics exist, be first to read them.

No spam, no tips, no promises.