Nandan Adhikari is a banker with a full-time job. The trading bots he built place and manage his trades in Indian stocks and options while he is at his desk. KANA Capital is where he writes down how that works.
The Indian market is open from 9:15 am to 3:30 pm, which is exactly when a salaried person cannot watch a screen. This is how the day is split between the bot and its builder.
Logs in to the broker, refreshes its access, and loads the instruments it will trade today.
On the way to the bank.
The market opens. It starts reading live prices and marks out the opening range.
At his desk, starting his actual job.
Looks for entries: a break out of the opening range, confirmed by trend and momentum indicators. Every order goes in with a stop-loss.
In meetings.
Manages open positions. It trails stops, books targets, and stays out when its signals disagree.
Still at work. The phone stays in his pocket.
Closes intraday positions before the bell.
Wrapping up the day's files.
The market closes. It writes every order, fill and price to the day's log.
Heading home.
Idle.
Reads the log. What fired, what didn't, and what went wrong.
Re-runs its backtests on the latest data so the next change is based on evidence.
Asleep.
Two programs, built separately so that a problem in one never touches the other.
Trades index options intraday. This is the fast one: options move quickly, so entries, stops and exits all have to happen without a human in the loop.
Trades shares in the cash market. Stocks move less in a day than options, so its targets are set from what a stock usually gives on an average day.
A trading bot is a set of written rules followed without exception. Each trade goes through the same five steps.
Live prices arrive from the broker's data feed, and the bot builds candles and indicators from them.
A signal counts only when several conditions agree. One indicator on its own is not enough.
Before any order, the bot knows how much it can lose on this trade and how much it has already lost today.
The order goes to the broker through its API. From then on the bot watches the stop-loss and the target.
The bot exits by rule, then records the real fill prices so the profit and loss figure is one you can trust.
It will follow a bad rule just as faithfully as a good one. The discipline is in the rules you write.
If exits are calculated from assumed prices instead of real fills, the results look better than they are. Record what actually traded.
History has no slippage, no delayed orders and no outages. Treat a backtest as a filter for bad ideas, not proof of a good one.
Run the bot on live prices with pretend orders for weeks. Most bugs show up here for free.
Logins expire, feeds drop and rate limits bite. A bot needs a plan for each of these before it holds an open position.
Automation does the trading. It does not do the reviewing, and that part still takes a person.
A banker during market hours, a trader because of the bots.
Nandan Adhikari works full time in banking and trades the Indian stock and options markets. Since he cannot sit in front of a screen from 9:15 to 3:30, he built bots that do the day-to-day trading for him.
KANA Capital is the name he trades and builds under. This site shares how the bots work and what he has learned, for anyone curious about automated trading.
Questions about how something was built are welcome. Requests for tips or trade calls will not get a reply.
Email shrikanacapital@gmail.com