StageTwo Momentum

Today, live from Kite

    2026 to the 13 Nov close

    The three model portfolios against the index

    These are rule-following model portfolios, not real accounts. Each has run by the same rules since 2010, so today's weights reflect winners that were left to grow.

    ₹10 lakh invested at the start of January 2026

    What the rules say to do next

    Head to head

    The three portfolios side by side

    Momentum v1Early-stageVCP, 30-week exit

    Return by month, 2026

    *November is to the 13th. Totals compound the months.

    MonthMom. v1Early-stageVCP 30-wkNifty 500Small 250
    Momentum v1Early-stage momentumVCP, 30-week exit

    What was bought and sold each month

    Which stock was held in which month

    Share of the portfolio at each month-end, in percent. A blank cell means the stock was not held.

    Model portfolio today

    Paper trade: the finalised strategies, from today

    Each strategy starts with ₹10,00,000 of virtual money on 17 November 2026 and follows its rules from here. No real orders are placed. This is the forward test: the only numbers on this dashboard that were not chosen after seeing the answer.

    Overall return so far

    PortfolioValue nowGainReturnInvested

    Each portfolio in detail

    Opening positions

    StockQtyEntry ₹Now ₹Stop-loss ₹Room to stopGainValue ₹

    The rules this portfolio follows

    Record

    Every buy, sell and rule check is added here with its date and price.

      What this test can and cannot tell you

      • Fills are today’s opening prices. Restricted-series stocks are priced on BSE, where Kite quotes them. Bhagyanagar India opened near its upper price limit with no sellers, so a real buy might not have filled.
      • Costs are charged. 0.2% a side, as in the backtest. Tax, dividends and slippage on thinly traded stocks are not.
      • A few months prove little. These strategies make most of their money from a handful of large winners, so any single quarter can look great or terrible. Judge the record after a full year at least.
      • The same stock can sit in two portfolios. Sona BLW, Shilpa Medicare and others appear in more than one, so the three are less independent than they look.
      • Trade-for-trade stocks (marked below) can be bought only for delivery and may hit price limits, so a real exit could take days.

      What to do next

      StockStatusStrength rankFri close ₹Now ₹Today

      Positions

      How it trades

      Growth of ₹1 lakh (log scale)

      Return by calendar year

      2026 month by month

      MonthMom. v1Early-stageNifty 500

      IdeaReturn/yr
      2010–18
      Return/yr
      2019–26
      Worst fallAvg hold
      sessions
      Trades
      a year

      What to do next

      StockStatusStrength rankFri close ₹Now ₹Today

      Positions

      How it trades

      Growth of ₹1 lakh (log scale)

      Return by calendar year

      2026 month by month

      MonthMom. v1Early-stageNifty 500

      IdeaReturn/yr
      2010–18
      Return/yr
      2019–26
      Worst fallAvg hold
      sessions
      Trades
      a year
      Every strategy tested so far, ranked on the unseen years

      Which strategy earned the most

      StrategyReturn/yr
      2019–26
      Return/yr
      2010–18
      Worst fallSince
      Jan 2024

      What each trader teaches, and what a backtest can check

      What each rule did when tested

      Average return per year across every variant with and without the rule, on the 1,000 most-traded stocks. These are mechanical tests on daily closing prices. The traders apply the same rules with intraday stops, discretion and fundamentals, which a backtest cannot copy.

      Rule2010–18
      without
      2010–18
      with
      2019–26
      without
      2019–26
      with

      Does shareholding data improve the strategies?

      Stock by stock: return over the next 12 months

      Median return of stocks that passed each test against those that failed. The first pair is every tradable stock, sampled monthly; the second is base-breakout signals only.

      TestAll stocks
      pass
      All stocks
      fail
      Breakouts
      pass
      Breakouts
      fail

      As an entry filter on the portfolios

      Return per year, 2020 to November 2026, when a stock must pass the test to be bought.

      Entry filterEarly-stageMomentum v2Leaders weekly

        What to track

        As a trader: all of this is in the backtests

          As an investor: none of this is in the backtests

            Sources

              All 19 strategies over 16 years

              Ranked by Sharpe ratio on the tuning years. A strategy earns trust when the unseen years hold up. Tap a row to load its charts, holdings and trades below.

              StrategyUniverseSharpe
              2010–18
              Sharpe
              2019–26
              CAGR
              2019–26
              Worst fall
              2019–26
              CAGR
              all years
              Worst fall
              all years
              Strategy detail

              Growth of ₹1 lakh (log scale)

              Fall from previous peak

              Return by calendar year

              Orders for the next session's open

              Watchlist: base breakouts in the last 45 sessions

              Every stock that fired the signal, whether or not the model portfolio had a free slot. Still valid means the price is above its exit average and stop.

              Signal dateStockPrice then ₹Close ₹SinceStatus

              Model portfolio holdings

              StockHeld sinceEntry ₹Close ₹P&LWeight

              Last 20 closed trades

              StockBoughtSoldDaysNet return

              Institutional changes, 17 November 2026

              Each change was tested on its own and together. Return per year from 2010, 2019 and 2024 to 13 November 2026, and the worst fall since 2010.

              Version2010–2019–2024–Worst fall
              Momentum v1
              Before34.6%45.5%45.2%-52.6%
              3 tranches only34.1%42.3%43.9%-52.1%
              15% cap only34.2%43.3%44.8%-55.0%
              Cap + no restricted series40.2%52.8%41.1%-52.9%
              Adopted: all three36.4%46.0%40.8%-52.1%
              Volatility target 20% (not adopted)30.4%35.3%33.3%-47.9%
              ₹1 cr/day liquidity floor (not adopted)35.5%47.8%33.0%-51.0%
              Early-stage
              Before (one rebalance calendar)33.2%42.5%37.9%-40.7%
              Same rules, calendar shifted 7 sessions27.9%35.0%26.4%-44.8%
              Shifted 14 sessions26.4%26.0%24.0%-48.3%
              No restricted series only29.7%33.9%23.1%-37.6%
              Adopted: 3 tranches + no restricted28.8%32.9%26.9%-43.0%
              ₹1 cr/day liquidity floor (not adopted)25.8%33.6%20.9%-45.0%
              Early + template
              Before31.0%41.9%34.7%-42.3%
              Calendar shifted 7 sessions22.1%29.6%20.8%-46.9%
              Shifted 14 sessions22.9%23.9%24.5%-46.3%
              Adopted: 3 tranches + no restricted25.7%30.8%26.7%-43.2%
              VCP, 30-week
              Before27.0%28.9%26.7%-31.4%
              Adopted: no restricted series26.4%28.0%22.4%-31.4%
              Rank breakouts by tightness (not adopted)27.6%28.4%22.4%-29.6%
              ₹1 cr/day liquidity floor (not adopted)19.4%25.7%20.2%-33.8%
              • The biggest finding: the early-stage numbers were partly luck of the calendar. Both rebalance every 21 sessions, and the original calendar happened to land on good days. Shifting it by a week or two cut early-stage from 33% to 26–28% a year, and early-stage + template from 31% to 22–23%. Splitting the money into three tranches a week apart removes that luck; the tranche figures are what to expect. Momentum v1 barely moved with the calendar (33.5–34.6%), so its result is sturdier.
              • Restricted-series stocks are now never bought. These trade-for-trade stocks can be bought only for delivery and often sit at their price limits, so a real portfolio cannot rely on getting in or out. History of the series comes from NSE bhavcopies to 23 September 2026, with Kite’s current list after that.
              • Momentum v1 now caps each holding at 15% of its tranche. With the other two changes, its worst calendar year improved from −34% to about −25% and the worst fall since 2019 from −48% to −37%.
              • Walk-forward check on early-stage + template. Choosing its stop (20/25/30%) and strength cut-off (70/80) each year using only earlier years returned 38.0% a year from 2012, against 38.2% for the fixed setting, so those two choices were not overfitted.
              • Not adopted: a volatility target (holding less when the portfolio is swinging hard) cut Momentum v1’s worst fall only a few points and cost about 4 points a year. A ₹1 crore/day liquidity floor cost 4–8 points a year: much of the edge is in thinly traded stocks.
              • Costs. Doubling trading costs to 0.4% a side lowered each strategy by about 1 to 1.5 points a year.
              • Idle cash belongs in a liquid fund: it adds about 0.5–1 point a year to early-stage and VCP. The figures above leave cash at zero return.

              How large these portfolios can get

              Share of past purchases where one position (a fifteenth of the portfolio) would have been more than 10% of the stock’s average daily traded value, a common institutional limit.

              Portfolio size₹1 crore₹10 crore₹50 crore
              Momentum v117.2%65.0%83.5%
              Early-stage11.9%58.8%80.3%
              Early + template9.2%57.0%78.3%
              VCP4.9%49.9%75.2%

              Up to about ₹1 crore these trade comfortably. Beyond a few crore, most purchases would move the price, and the liquidity floor (and its lower returns) would be needed.

              What was corrected, and what did not work

                Why momentum missed 2024–26

                  Return by year

                  Your two test cases

                  StockBase-breakout signalPrice then ₹SinceMomentum (Nifty 500) first bought

                  Prices are adjusted for later splits and bonuses. Cupid traded about ₹1 crore a day in September 2023, outside the 1,000 most-traded stocks, so no strategy here was allowed to buy it then. Early-stage momentum bought Cupid on 19 August 2025 at ₹34 and still holds it; it never held Laurus Labs, which did not rank in its top 15 after either breakout.

                  How good is an early signal on its own?

                  How hard early-stage momentum was pushed

                  TestCAGR 2019–26 %CAGR all years %Worst fall %

                  Read this before trading it

                  • Every version falls hard. Peak-to-trough losses run from 41% to 49%, and each one lost money in 2011, 2018 and 2025. The corrections fix where the 2024–26 gap came from; they do not remove losing years.
                  • The 35% figure is fragile. Early-stage momentum made 28% to 37% a year in the unseen period depending only on which day of the month it rebalanced. Plan on the lower end, before the data bias below.
                  • An early signal is a candidate, not a verdict. Price and volume cannot see debt reduction, margin expansion or a capex cycle ending. Half of all base breakouts go nowhere; the portfolio result comes from cutting those quickly and holding the few that run.
                  • Returns are overstated by missing dead stocks. The price file drops companies a year after delisting. Roughly 15% of 2014 Nifty 500 members and 5% of 2020 members have no data. The gap is wider among the smaller stocks the broad-market strategies trade, so trust their numbers less.
                  • Smaller stocks cost more to trade. The same per side is assumed everywhere. Stocks near the bottom of the 1,000 trade ₹1–2 crore a day, so real slippage will be higher and position sizes are limited.
                  • Dividends are excluded from strategies and benchmarks alike. Short-term capital gains tax is not deducted.
                  • Before 2014 the Nifty 500 universe is the 500 most-traded stocks, because index membership records start in 2014.
                  • VCP’s exit changed on 17 November 2026 from a daily close below the 50-day average with a 12% stop, to a weekly close below the 30-week average with a 20% stop. The old version stays on the Record tab for comparison. Of the exits tried, the 200-day daily average did slightly better over 16 years (29.2% a year, worst fall −31%); the 30-week one was chosen for its stronger 2024–26 record and weekly checking.
                  • All three switched to the 1,250 most-traded stocks on 17 November 2026. That roughly means companies worth ₹1,000 crore and more. The switch was made after seeing that it helped in 2019–26; before 2019 it changed nothing, because few small stocks traded enough. Momentum v1’s worst fall deepens to about −53% (it was −45% on the Nifty 500), and the robustness checks on the Record tab were run on the earlier 1,000-stock list. Missing data on delisted companies flatters the smallest stocks most, so this list is the one most exposed to that bias.
                  • Early-stage + template was dropped from the core strategies and the paper trade on 17 November 2026. It added little over plain early-stage momentum, held mostly the same stocks, and was the most dependent on its rebalance calendar. It stays on the Record tab for reference.
                  • The 25% stop on early-stage + template was picked after testing 20%, 25% and 30%. The three gave 28.9%, 30.6% and 28.9% a year and win rates of 42.9%, 43.8% and 44.9%, so the 25% result is partly luck. The same change on plain early-stage momentum lowered its return from 30.6% to 29.2%. Expect about 29%, and watch the paper-trade record.
                  • Sixteen years is one market history. variants were compared, so some of the edge is luck. Treat the test-period numbers as an upper bound.