Log in
Account
Account
Tools

Method and FAQ

Every model here is tested on daily NSE prices from 2015, on the 500 most traded stocks each month, after delivery charges and slippage. A backtest, that is a replay on past prices, is only as honest as these assumptions.

Stock list: which stocks can a test hold?

Each month we rebuild the stock list: the 500 NSE stocks with the most money traded in the past six months, from about 2,000. A stock counts only in months it was on the list.

  • Stocks removed from the exchange (delisted) are missing from our price data, which makes results look better than they would have been.
  • We have no month-by-month company-size data, so money traded stands in for size.
  • Sectors follow the NIFTY Total Market industry list. A stock outside that list goes in the industry whose past-year returns it tracks most closely, using only data up to that day. The sector list is one saved snapshot applied back to 2015. Placing stocks outside it is a best guess, so sector results are less certain.
Stocks
StocksEach month the 500 most traded NSE stocks, from about 2,000, become the list the model may hold.About 2,000 NSE stocksTop 500 by traded value over the last six monthsOnly these can be held that monthRebuilt every month · delisted stocks are missing
StocksEach month the 500 most traded NSE stocks, from about 2,000, become the list the model may hold.About 2,000 NSE stocksTop 500 by traded valueover the last six monthsOnly these can be heldthat monthRebuilt monthly · delisted missing

Sketch. A stock counts only in the months it was on the list.

How is our backtest different from a typical one?

  • Which stocks could it have held?

    Typical backtestToday’s index list, applied to the past
    TickjuryThe top 500 by traded value as it stood each month
  • Stocks removed from the exchange?

    Typical backtestQuietly dropped
    TickjuryMissing from our data, and we say so
  • What did trading cost?

    Typical backtestOften ignored
    TickjurySTT, stamp duty, exchange fees and GST on every trade, plus worse prices
  • When did a trade start?

    Typical backtestAt the close that triggered the rule
    TickjuryAt the next day’s open
  • Was it one lucky stock?

    Typical backtestNot checked
    TickjuryRerun without the top 5 winners
  • Could random picks have done as well?

    Typical backtestNot checked
    TickjuryFor the Robust model: random picks, same stake size and costs
  • Were the settings just lucky?

    Typical backtestThe best of many, shown as if it were the only one
    TickjuryFor the Robust model: chance of overfitting, luck-adjusted score
  • Who picks the stocks?

    Typical backtestSomeone else’s list
    TickjuryThe model’s own rules. We publish tested models and give no advice.

Common questions

What data do you use?

We use daily prices of NSE stocks from 2015, and tests start in 2016. Each month the stock list is the 500 stocks with the most money traded at that time, out of about 2,000. Stocks later removed from the exchange are missing from our data.

Are trading costs included?

Yes, every trade pays the standard delivery charges: STT, stamp duty, exchange and SEBI fees, GST and the demat charge. It also pays 0.1% slippage each way, meaning a slightly worse price, at the next day’s open. We also rerun the research models at 0.3% slippage. Tax on profits is not included.

Can it trade for me?

No, Tickjury places no orders and cannot connect to your broker account. It only runs backtests, which are replays on past prices.

What does it cost?

Research, Learn, the open studies and the tool pages are free. A locked research model's trade history and holdings open with credits for 30 days; every account gets free credits each month, and packs start at ₹100. Two proprietary models, Rowan and Teak, stay closed. See Credits.

What does the site keep about me?

The server keeps a request log (address and page). The public site reaches the server through a secure tunnel (Cloudflare). The site uses cookies only to keep you logged in and to protect your account. It loads no analytics or tracking scripts.