☰
Trading Operationsvictorngoma.com

Hi there · Backtests · what the tests say

Backtests

How each strategy did on past prices, the newest first. Clear results, including the ideas that did not work.

ⓘA backtest is a replay on past prices. It shows what would have happened, not what will happen. Every study is checked on years the rules had not seen. Updated 15 Nov 2026.
Studies since October35
Became a paper strategy15
No edge found15
New strategies8

Backtests of the new strategies

One card for each strategy started this month. Each table shows the years the rules were built on, then the later years used to check them.

MES 2H RSI v.16.11.2026Futures

Buys one S&P 500 micro future on a night dip while the trend is up. Sells at a set profit price, or when the bounce is over.

MeasureBuilt on 2019–2024Checked on 2025–2026
Trades15151
Winners83 %82 %
Profit factor3.043.65
Sharpe (daily)1.912.61
Profit a year, one contract$1,498$2,687
Biggest drop from a high−$1,002−$789
Worst trade−$1,002−$671
Deepest open loss−$2,180−$989

The first version on the same test: 72 % / 67 % winners, profit factor 1.51 / 1.85, biggest drop about −$3,000.

Data: 1-minute S&P 500 futures bars, 2019 to 2026. Contract changes are closed, not counted as profit.  ·  Size: One contract

Keep in mind: No stop-loss, and about 25 trades a year. Sharpe is under 2.5 in the tuning years. Part of the profit is the market’s own rise.

M2K 3H RSI v.16.11.2026Futures

Buys one Russell 2000 micro future on an evening dip while the trend is up. Sells at a set profit price, or when the bounce is over.

MeasureBuilt on 2019–2024Checked on 2025–2026
Trades15561
Winners76 %77 %
Profit factor2.612.29
Sharpe (daily)1.901.79
Profit a year, one contract$785$1,271
Biggest drop from a high−$378−$590
Worst trade−$237−$459
Deepest open loss−$524−$675

The first version on the same test: 68 % / 73 % winners, profit factor 1.35 / 2.45, biggest drop −$1,625.

Data: 1-minute Russell 2000 futures bars, 2019 to 2026. Contract changes are closed, not counted as profit.  ·  Size: One contract

Keep in mind: No stop-loss. Winners are under 80 % and Sharpe is under 2.5. A steadier version of the first rule, not a bigger earner.

Ladder Pot v.15.11.2026ETFs

Buys sharp dips in seven 3x funds with small lots. Never more than $5,000 in use. Every lot is out within 3 days.

MeasureBuilt on 2022–2024Checked on 2025–2026Last 12 months
Steadiness (Sharpe)3.383.042.94
Winning trades96.6 %95.7 %94.3 %
Profit factor5.876.674.31
Profit per year$1,115$858$818
Worst day−$103−$103−$103
Biggest drop from a high−$117−$117−$117

Worst day and biggest drop are for the whole test, 2022–2026. Each single year: Sharpe 4.68, 2.40, 3.03, 2.90, 3.20.

Data: 1-minute prices, 2022 to 2026. 1,661 trades.  ·  Size: $5,000 paper account

Keep in mind: Two of the ten rules (UDOW and FAS) were chosen after seeing 2025–2026. Paper trading is the real test.

ETF Basket v.15.11.2026ETFs

20 sector funds, each bought on a dip inside an up-trend. New: every lot is sold by day 5 and at most $5,000 is in use.

MeasureBuilt on 2022–2024Checked on 2025–2026Last 12 months
Steadiness (Sharpe), $5,000 limit3.862.981.33
Winning trades86 %80 %74 %
Profit factor3.52.41.6
Sharpe together with the Ladder Pot4.793.742.25
Profit per year together with the Ladder Pot$3,091$2,545$1,606

Winning trades and profit factor were measured with a $10,000 limit. Worst day of the two books together: −$164.

Data: 15-minute to 1-hour prices on 20 funds, 2022 to 2026.  ·  Size: $5,000 paper account

Keep in mind: The last 12 months were a calm, rising market and are the weakest period. The 5-day rule gives up some winners to avoid stuck lots.

SPX Harvester Capped v.13.11.2026Options

The SPX same-day option seller with one hard rule: the most it can lose in one day is $1,500.

MeasureBuilt on 2022–2024Checked on 2025–2026
Winning trades (all legs)83 %80 %
Profit factor (all legs)3.082.19
Profit factor (main legs only)2.522.01

Whole test: about $25,500 a year with all legs ($16,300 with the main legs only), worst day −$1,480, biggest drop from a high −$4,190.

Data: Real option quotes, 1-minute, 2022 to October 2026.  ·  Size: Sized for a small account: $1,500 at risk a day

Keep in mind: The final set of rules was chosen after seeing 2025–2026. The first version of this Harvester made much less in live paper trading than in its test. Paper is the judge.

SPX Quiet Condor v.12.11.2026Options

On a quiet afternoon it sells one same-day SPX iron condor at 2:30 PM and holds it to the 3:00 PM close.

MeasureBuilt on 2022–2024Checked on 2025–2026
Trade days7469
Profit factor2.033.00
Total profit, one condor$7,556$10,823
Biggest drop from a high−$2,089−$1,118

Data: Real option quotes and 15-minute futures bars, 2022 to 2026.  ·  Size: One condor a day at most

Keep in mind: The quiet test was first seen on the whole sample. If real fills are 30 % worse than the test, the edge gets thin. Paper test runs about 8 weeks.

M2K Morning Trend v.12.11.2026Futures

Buys one small-cap micro future when the morning trend turns up, and sells when the trend turns down.

MeasureBuilt on (older years)Checked on (recent years)
Profit factor, no filter1.400.98
Profit factor, with the calm filter1.951.21
Profit factor, + daily strength filter (recorded, not used yet)2.351.33

The filter lifts every period, but the recent years are only a little above break-even.

Data: 15-minute M2K futures bars, 8 years.  ·  Size: One contract

Keep in mind: The smallest edge of the new strategies. It runs on paper to see if the filter keeps working.

PreMarket Book v.14.11.2026Stocks

Five small stock trades picked from the 8:28 AM pre-market scan. Three buy, two sell short. $500 at risk a day.

MeasureBuilt on 2022–2024Checked on 2025–2026
Winning trades, all five86 %84 %
Profit factor, all five3.472.48
Winning trades, the three buy rules87 %85 %
Profit factor, the three buy rules3.732.62
Winning trades, the two short rules84 %83 %
Profit factor, the two short rules2.802.21

Whole test: about $1,400 to $1,800 a year, worst day −$188.

Data: 1-minute prices on about 2,300 stocks, April 2022 to October 2026 (1,135 days).  ·  Size: $5,000 paper account

Keep in mind: The rules were chosen after seeing 2025–2026. Losing trades are bigger than winning trades, so the high win rate matters. Short sales need a margin account.

How to read the numbers

Winning tradesOut of 100 trades, how many made money.
Profit factorDollars won for each dollar lost. Above 1.5 is good. Above 2 is strong.
Steadiness (Sharpe)How smooth the profits are from day to day. Above 2 is very steady.
Worst dayThe biggest loss in one single day of the test.
Biggest drop from a highThe largest fall of the account from a peak before it recovered.
Built on / Checked onThe rules are made on the first years, then checked on later years they had not seen.

All studies, newest first

Each line is one question, the answer in one sentence, and what was done with it.

DateStudyGroupResultOutcome
2026-11-16The four futures strategies, deeper searchCan each futures strategy reach 80 % winners, profit factor 2 and Sharpe 2.5?FuturesNone reaches all three with a sound rule. Two steadier versions were found and now run on paper: MES 2H RSI and M2K 3H RSI.Built on paper
2026-11-15Ladder Pot, round 3Can the dip ladder stay steady in every period, crashes included?ETFsYes, with a crash switch and seven funds. Sharpe 3.4 / 3.0 / 2.9, worst day −$103.Built on paper
2026-11-14The six ETF strategies, honest numbersWhat do the ETF strategies really make when losing lots are counted every day?ETFsOnly the ETF Basket was truly strong. A 5-day rule and a $5,000 limit keep it steady.Built on paper
2026-11-14Pre-market rules, high win rateCan pre-market stock rules win more than 75 % of the time with a profit factor above 2?StocksFive small rules together: 84 % winners, profit factor 2.5, worst day −$188.Built on paper
2026-11-14Failed-bounce spreadsDoes a failed bounce on a weak day pay as a call spread?OptionsYes, as two small legs inside the $1,500 day budget. Profit factor 2.2 in the check period.Built on paper
2026-11-13Harvester review and day capCan the Harvester earn more and never lose more than $1,500 in a day?OptionsYes. Profit factor 2.5 / 2.0, about $16,300 a year, worst day −$1,480. Stops and early exits did not work as a cap.Built on paper
2026-11-13Trend line + RSI on M2KDo three trend-line gates with RSI entries work on M2K, 1 minute to daily?FuturesNo. 2,800 combinations, two separate runs, none real. 1 to 5 minute charts always lose.No edge
2026-11-13ATR trailing stop on M2KDoes an ATR trailing stop pay on M2K, 5 minutes to daily?FuturesNo. 28 settings, long and short: none passed.No edge
2026-11-12Quiet-afternoon condorIs the afternoon condor better on quiet days?OptionsYes. Profit factor 2.0 / 3.0 when the market is squeezed and the day before was normal.Built on paper
2026-11-12M2K morning trendDoes a calm-market filter improve the morning trend rule?FuturesYes. Profit factor 1.40 / 0.98 becomes 1.95 / 1.21.Built on paper
2026-11-12Danger-day flagsCan fear gauges, trend and event flags tell which days to skip?OptionsNo. They forecast how big the move is, not its direction. 0 of 20 skip filters helped.No edge
2026-11-12Squeeze as a trade signalDoes a price squeeze work as a trend or a skip signal?OptionsNot as a signal. Only use: a squeeze at 2:15 PM means a better condor day (used in the Quiet Condor).Lesson learned
2026-11-12Weekly income on $5,000Can a $5,000 account pay a steady weekly income?Options12 ideas tested, none passed. Realistic: about $50 to $100 a week on average, with losing weeks.No edge
2026-11-05Days when the S&P is up 1.6 %Is there a trade after the index is up 1.6 % in the day?OptionsNo. It happens on 8 to 10 % of days and does not fade. Selling call spreads there breaks even.No edge
2026-11-03MES and M2K, 1 hour to dailyDo oversold rules work on index micro futures?FuturesYes, overnight only. MES 2-hour: profit factor 1.70, every year positive. M2K 3-hour: 1.68.Built on paper
2026-11-02Overnight futuresIs there a pattern in the overnight session?FuturesOnly one small one: buy Sunday evening after a red Friday, about $470 a year per contract.Built on paper
2026-11-01News days: fade the band breakDoes fading a band break pay on Fed, inflation and jobs days?ETFsNo, it is a coin flip. One small lead: short the afternoon down-break on Fed days.No edge
2026-11-01Trend-skewed condorDoes a condor leaned with the first 5 to 15 minutes pay?OptionsNo. 0 of 3,456 combinations paid on real quotes.No edge
2026-10-31Big one-day drops in mega-capsIs a 10 % one-day drop in a giant stock a buy?StocksNo edge. 128 cases: the bounce is the same as any normal day.No edge
2026-10-31Oversold gate, every timeframeWhich chart is best to call the market oversold?Options4-hour for SPX spreads (profit factor 3.76, 192 trades). Under 2 hours it is random.Built on paper
2026-10-30Four old SPX trend rules, againDo the four retired SPX trend rules work on real quotes?OptionsNo. The 15-minute trend signal is a coin flip. They stay retired.No edge
2026-10-28Trend gate + RSI ladder on 38 fundsWhich funds pay with a dip ladder inside a trend gate?ETFsA basket of 20 funds, each on its own rule: steady and small drawdown. Became the ETF Basket.Built on paper
2026-10-27Gold fund ladderDoes the ladder work on the gold fund?ETFsThree small survivors out of 867,000 runs. Kept as a lead, not built.Lead, not built
2026-10-2617 published ETF patternsDo well-known calendar and gap patterns still work on 47 funds?ETFsThree pass (turn of the month, emerging markets after a US drop, gap-downs). Small: about $2,000 a year on $15,000.Lead, not built
2026-10-25One ladder rule for 47 fundsDoes one family rule work across all funds?ETFsNo. It fails the last 12 months against random entries and against buy-and-hold.No edge
2026-10-23Band break put spreadsDoes a close under the daily band pay as a put spread?OptionsYes: 47 trades, 91.5 % winners, profit factor 3.74. Became SPX Band.Built on paper
2026-10-22Harvester and condor on other fundsDo the SPX rules work on QQQ, IWM, TQQQ and TNA options?OptionsMostly no. One lead: the QQQ same-day bear leg (90 % winners on 41 trades).Lead, not built
2026-10-20Does RSI call the turn?At which chart does an extreme RSI reliably reverse?ETFsNone. 40 to 54 % against a normal 45 to 48 %. Oversold TQQQ on the daily chart keeps falling.No edge
2026-10-11Share-only morning scannerCan a pre-market scanner pick stocks to buy, with no options?StocksNo. Five ideas on 2,338 stocks failed the check period.No edge
2026-10-11Liquidity firstDo cheap, liquid calls on strong stocks pay?OptionsNo. The price gap between buy and sell (20 to 47 %) eats the edge.No edge
2026-10-09Two-day oversold bounceDo stocks bounce after two hard down days?StocksNo edge on 552,000 cases. Every rule was equal to random.No edge
2026-10-05Whole-market ladder scanCan the ladder idea pick stocks each morning?StocksLong side yes (profit factor 2.64 against random +$1,000). Short side fails.Built on paper
2026-10-04175 published option rules on real quotesWhich published option rules survive real prices?OptionsOne: the call diagonal on index funds after an oversold day (profit factor 3.8). Single stocks and earnings rules lose.Built on paper
2026-10-03Ladder gates, exits and shortsCan a trend gate, a faster exit or a short side improve the ladders?ETFsA gate cuts drawdown 80 to 95 % but also most of the profit. Faster exits are one dial. The short side pays only in 2022.Lesson learned
2026-09-30Harvester early exitsDoes any early exit beat holding to the close?OptionsNo. Targets, time exits and stops all make less than holding to settlement.No edge
Earlier studies, May to September 2026

Kept for the record. Several of these strategies have since been stopped or replaced.

StudyScopeDateBasisKey resultStatus
CCThe Condor CalendarSPX 0DTE2026-09-06Real quotesThe live 15:30 condor (0.9 straddles / 25 wide / 0.35 side stop) re-priced on 871 sessions of real quotes, winters included, split by side and by day. As traded PF 1.16. Wednesday 1.54 and Friday 1.88 earn in both halves; Monday 0.89 and Tuesday 0.86 lose in both; the 1st–7th loses. Book deployed 2026-09-08: one contract every day plus a second on Wed/Thu/Fri inside the 8th–21st: 1,117 lots, PF 1.30, +$44.9k vs +$20.8k, worst day −$1,926, 2024 still negative.✓ Validated
HCThe Harvester CalendarSPX 0DTE2026-09-06Real quotesDual Harvester v2 re-priced on 973 sessions of real SPXW quotes (2022–26, winter afternoons filled): as traded PF 1.05. Day of month is the edge: 15th–21st PF 1.97 (train 2.35 / test 1.67), 1st–7th PF 0.75 in both halves. Per-leg weekday gates + skip 1st–7th + one entry per leg per day + a 2nd lot on the strongest days: 543 lots, 78% win, PF 1.76, Sharpe 3.8, worst day −$4,240. Deployed paper and live 2026-09-08.✓ Validated
15The 15:30 CondorSPX 0DTE2026-08-28Real quotes1,008 sessions, 19 M combinations. 4 of 9 design goals met at a half-spread fill: CAGR 29.4%, drawdown 7.0%, Sharpe 2.28, PF 1.74, 69.7% wins. Profit factor rises monotonically with the hour — morning entries lose. PF 2.2 at a 75–85% win rate is arithmetically unavailable. Test half beat train (PF 1.90 vs 1.59). Not deployed.! Caveat
0CThe 0DTE CeilingSPX 0DTE2026-08-25Real quotes1,006 sessions. Profit factor 2 unreachable at any strike, width, entry, exit or trigger tested. Screening rules on the first 604 sessions gave a train→test correlation of −0.11.– Negative result
DH0DTE Dual Harvester — real-quote buildSPX 0DTE2026-08-22Real quotes476 sessions. Long leg 81.5% win / PF 1.27; short leg 71.9% / PF 1.27. Long leg reached the 98th percentile against 200 random entry sets.✓ Validated
MT1DTE multi-timeframe VIDYA/RSISPX 1DTE2026-08-22Real quotesModelled PF 9.31 fell to 2.86 on real fills — a 30-day VIX used as 1-day vol. Shipped at 2 concurrent sequences: PF 2.11.! Caveat
PSPremium-stop measurementSPX 0DTE2026-08-22Real quotesThe 2.5× credit stop turned the long leg from PF 1.21 into 0.95. Fires on 39.5% of trades and cuts winners that recover by settlement. Removed from that rail.– Negative result
BB1DTE Bollinger violation v2SPX 1DTE (bull)2026-05-16ModelSharpe surface monotonic across a 4×4 grid (ridge 4.7–5.7), so not a fitted peak. But dormant in 87.6% of months — 37 trades in 11 years.! Caveat
RR7DTE RSI/RSI + server-side stopSPX 7DTE2026-05-04Real quotesA $30 stop lifted Sharpe from 0.947 to 1.843 with the win rate unchanged at 72.4%. Rests on 29 trades.✓ Validated
1D1DTE vs 0DTE end-of-day holdSPX 0DTE + 1DTE2026-05-06Real quotes1DTE ITM30/W10 Sharpe 6.57 and ITM20/W10 6.53, against the 0DTE baseline at 2.15. 225–226 trades each.✓ Validated
TS176-trade settlement studySPX 0DTE + 1DTE2026-07-30Real quotesWith-trend entries +$5.7k against counter-trend −$14.6k; in-the-money entries −$12.2k. Directly shaped the live trend gate.✓ Validated
TFHarvester timeframe robustnessSPX 0DTE (v3)2026-07-30ModelPositive on all six timeframes tested. The 5-minute chart was selected — a plateau, not a peak.✓ Validated
1M1-minute vs 15-minute entry detectionSPX 0DTE2026-05-04Real quotesSharpe gap +1.141 in favour of the earlier trigger, clearing the +0.3 threshold. Entry moved to 1-minute.✓ Validated
SVStrategy validation 2026SPX 0DTE + 1DTE2026-05-13Real quotesNo production changes recommended. Both candidate improvements landed inside the ±20% inconclusive band. Walk-forward passed; no overfit flags.✓ Validated
SWSPY signal — eight structures head to headSPY options2026-08-20Real quotesSame 134 signals in 8 wrappers. Long call +24.2% per trade; the 2%-OTM credit spread won 89.7% of the time and still finished at PF 0.95.✓ Validated
EOOversold rules across 25 ETFs25 sector & asset ETFs2026-07-08Market prices2,267 trades over 16.5 years. All 25 funds clear break-even — and 20 of 25 still return less than buying and holding the same fund. Wins on drawdown and exposure, not return.! Caveat
ECETF credit spreads on the same signals8 underlyings2026-07-09Real quotesNo trustworthy edge. Not one of 85 combinations reached 30 trades under pessimistic fills. Oversold days and rich premium rarely coincide — 50–87% of entries were filtered out.– Negative result
TRTrend-flip study — VIDYA and SupertrendSPY, 6 timeframes2026-08-20Market prices1,248 combinations. The long side works and improves as the timeframe slows; the short side failed on every timeframe — 6 of 292 setups profitable, all six noise.! Caveat
PBVIDYA + RSI pullback — entry variantSPY daily2026-08-20Market pricesBuying the fall beats buying the bounce: +0.484% per trade at t = 6.43 over 190 matched pairs, with identical maximum adverse excursion.✓ Validated
ST157-name dip engine — inherited validation157 US stocks2026-08-14Market pricesNo direct backtest exists on these 157 names. Each stock inherits the rule validated on its sector fund (2,267 trades, 16.5 years). Volatility, single-name shocks and survivorship all bias live results the same way — downward.! Caveat
M2M2K double-RSI short as a hedgeMicro Russell futures2026-08-16Market pricesNot crash insurance. The rule needs an overbought market to arm, so it produced 0 signals in the 2018 Q4 selloff and 0 in the 2020 crash. It harvested +$1,128 in the 2022 grinding bear — the only regime it suits.– Negative result

Results are measured on past prices and are not a promise of future results. Option studies use real quoted prices and pay the gap between the buy and sell price on every trade.