SD1H4H — Recommendations

← Markets Hub · Strategy Priority Ranking · Best Validated Version · All Analyses · one pick per symbol, ranked

Recommendations

Methodology: for each of the 6 markets, this page starts from the single (structure engine, strategy, validated filter/mechanic if any) combo with the highest risk-adjusted score among combos that survive real trading costs -- the same score used throughout this site: (Monte Carlo probability of profit − Monte Carlo risk of ruin) × a confidence factor scaled down for samples under 200 trades, deliberately the SAME formula as Strategy Priority Ranking and Best Validated Version. That score alone is not enough to earn a live "Recommended" badge, though (added 2026-08-26) -- a combo also has to trade often enough that a realistic losing streak resolves in weeks, not months, and have enough total trades (this project's own 200-trade Tier 1 bar) that its own streak/win-rate statistics are themselves trustworthy. See the Trade Frequency table below for exactly why this matters: Monte Carlo probability-of-profit describes a long-run simulated outcome, not how long a real losing streak actually lasts in calendar time -- a combo can look strong on the score alone and still be impractical to actually hold through. Three statuses result: Recommended clears both bars; Track — Not Yet Tradeable has a positive, cost-surviving edge but trades too infrequently / too thin a sample to trade responsibly today; No Recommendation means no combo on that market survives costs at all (S&P 500). Cards are ordered Recommended, then Track, then No Recommendation, by score within each tier -- this page is a ranked list, not 6 independent write-ups in an arbitrary order. Where a validated finding exists for a market/strategy (an entry filter, stop mechanic, or sizing scheme -- see All Analyses), it's compared on equal footing against every unfiltered baseline and picked only if it actually scores higher, not automatically preferred. Every card, including Track-only ones, includes an Implementation section: the exact entry/stop/exit rule as literal, ordered steps -- precise enough to actually apply on a real chart, not just a backtest description (Track cards say plainly that it's reference material, not a live call). All picks are default 1H/4H pair, current as of 2026-08-26's data; not a guarantee of future performance, and not a portfolio-level recommendation -- cross-market correlation/allocation is explicitly out of scope (see All Analyses, H15).

Trade Frequency & Losing-Streak Reality Check

Monte Carlo probability-of-profit and risk-of-ruin figures (used for the score above) describe an account's SIMULATED long-run outcome -- they say nothing about how long a real losing streak actually lasts in calendar time, which is what a trader has to sit through to reach that long run. The table below is that reality check, computed directly from each combo's own trade history (2026-08-26): trades/year, raw win rate (not Monte Carlo probability of profit -- a different, trade-level number), the longest losing streak actually observed, and roughly how many real-world days that streak spanned at the combo's own trading pace. This is the basis for downgrading Bitcoin, EUR/USD, and Silver from live recommendations to "track only" below -- their Monte Carlo numbers look reasonable in isolation, but 18-22 trades/year means their worst historical losing streaks translate into 4.5 to roughly 12 months of consecutive real-time losses, long enough that a trader has no reliable way to distinguish "the edge is fine, this is just variance" from "the edge quietly broke" until it's already over. Gold and Nasdaq trade often enough (every 1-2 days) that the same kind of losing streak resolves in weeks.
MarketTotal TradesTrades / YearWin RateWorst Losing Streak~Calendar Time to Resolve
Gold (B)744161.323.9%15 trades~35 days
Nasdaq (A)1,170253.033.4%12 trades~17 days
Bitcoin (A)9921.921.2%12 trades~200 days (6.5 months)
EUR/USD (A)8519.025.9%7 trades~134 days (4.5 months)
Silver (A)8118.324.7%18 trades~358 days (~1 year)
#1Gold (XAU/USD)Recommended
Original engine · Strategy B · ENGINE=lookback_10 (10-bar stop lookback, vs. 30 baseline)
744 tradesnet avg +0.213Rnet total +158.6Rprob. profit 92.7%risk of ruin 4.6%median DD 40.6%
SummaryThe single highest risk-adjusted score on the entire site. Shortening the swing-based stop's lookback window from 30 to 10 bars improves the already-strong Gold Strategy B baseline, net of costs, and held up 3/3 walk-forward folds.
StrengthsLargest sample among every market's top pick (744 trades, well past the 200-trade reliability bar) combined with the lowest risk of ruin of any recommendation on this page. A genuine stop-mechanism change, not a curve-fit entry filter, and walk-forward validated rather than a single-split result. Trades roughly every 2 days (~161/year) -- see the Trade Frequency table below -- so even its worst historical losing streak (15 trades) resolves in about 5 weeks, not months.
WeaknessesEdge is regime-dependent, not uniform across market conditions -- gold's base Strategy B card found the edge concentrated in 2023-2025's sustained bull run, with a flat-to-negative 2022-mid 2023 stretch blended into the total. Median drawdown is still meaningful (~40%) even on the best pick on the site -- expect real equity-curve variance, not a smooth curve.
ConsiderationsDoes not stack with the no_event_bias mechanic (also Strategy B) -- pick one, not both. Gold Strategy A with the Tokyo/Asian session filter (net avg R +0.142, 85.3% probability of profit, ranked #2 overall) is a legitimate parallel stream to run alongside this pick, not a replacement for it -- different entry logic and trade timing, closer to a diversifying second position than a competing choice.
Implementation — step by step
  1. Pull up two charts for Gold (XAU/USD): a 4H chart (for bias) and a 1H chart (for entries/exits) -- both need to stay in sync, so run them on the same feed.
  2. On BOTH charts, track structure breaks (BOS/CHoCH) the same way: Swing highs/lows are confirmed by a simple 2-candle reversal: an up-candle immediately followed by a down-candle marks a confirmed swing high at the down-candle's high; a down-candle immediately followed by an up-candle marks a confirmed swing low at the up-candle's low. The prevailing bias flips to bullish the moment a newly-confirmed swing high exceeds the previous one, and to bearish the moment a newly-confirmed swing low undercuts the previous one.
  3. Every break requires the breaking candle's body to close past the level AND to have closed past the prior candle's own wick on the same side first (wick-confirmation) -- a close that clears the level but not that wick doesn't count until a later candle confirms it. A break in the SAME direction as the prevailing bias is tagged BOS (continuation); a break AGAINST the prevailing bias is tagged CHoCH (reversal).
  4. On the 4H chart, note the current bias (bullish or bearish) from the last confirmed 4H break, per steps 2-3.
  5. On the 1H chart, wait for a 1H break that is EITHER (a) a bullish/bearish BOS while the 4H bias already agrees (a continuation entry), OR (b) a bullish/bearish CHoCH while the 4H bias disagrees (a reversal entry) -- AND the 1H candle's close is above its own 50-period EMA for a long, or below it for a short (mandatory on every market, both strategies). Enter at that 1H candle's close.
  6. Place the stop at the low (long) or high (short) of the most recent OPPOSITE-colored 1H candle within the last 10 bars before entry -- i.e. the last candle that would have signaled the other direction. Skip the trade entirely if no such candle exists in that window, or if the resulting stop would already be on the wrong side of entry.
  7. Hold the position until the stop from step 5 is hit, OR until a newly-CLOSED 4H candle prints an opposing BOS or CHoCH -- ignore any 1H-level opposing signal in the meantime; this is a deliberately slower, higher-timeframe trailing exit, not a typo.
#2Nasdaq 100 (US Tech 100)Recommended
Original engine · Strategy A (unfiltered baseline)
1,170 tradesnet avg +0.053Rnet total +61.7Rprob. profit 74.0%risk of ruin 10.2%median DD 42.0%
SummaryThe clearest non-gold recommendation on the site, and the only one where the plain unfiltered baseline is already the best available combo -- second-highest score on the entire site, behind only gold's top pick.
StrengthsLargest, most reliable sample among every market's recommendation (1,170 trades). The 2026-08-26 mechanics rollout tested all 3 gold-derived mechanics here too and none improved on this baseline for its pre-registered pairing (see All Analyses) -- this pick has been actively stress-tested against alternatives, not just left untouched. Also the HIGHEST-frequency pick on this page (~253 trades/year, roughly one every business day, per the Trade Frequency table below) with the best win rate of the five (33.4%) -- its worst historical losing streak (12 trades) resolves in about 2.5 weeks.
WeaknessesThinnest edge (+0.053R/trade) of every market recommended here -- relies on a large trade count to compound to a meaningful total, not a large per-trade edge. Median drawdown is high (42.0%) relative to that edge size.
ConsiderationsStrategy B on the same market is materially worse (barely survives costs, near-coin-flip Monte Carlo profile) -- unlike gold, don't run both strategies here expecting genuine diversification. trail_chop_1h (trailing 1H structure-flip count) is flagged elsewhere on this site as walk-forward robust but not yet turned into an actual entry filter -- the natural next research step if pursuing Nasdaq further.
Implementation — step by step
  1. Pull up two charts for Nasdaq 100 (US Tech 100): a 4H chart (for bias) and a 1H chart (for entries/exits) -- both need to stay in sync, so run them on the same feed.
  2. On BOTH charts, track structure breaks (BOS/CHoCH) the same way: Swing highs/lows are confirmed by a simple 2-candle reversal: an up-candle immediately followed by a down-candle marks a confirmed swing high at the down-candle's high; a down-candle immediately followed by an up-candle marks a confirmed swing low at the up-candle's low. The prevailing bias flips to bullish the moment a newly-confirmed swing high exceeds the previous one, and to bearish the moment a newly-confirmed swing low undercuts the previous one.
  3. Every break requires the breaking candle's body to close past the level AND to have closed past the prior candle's own wick on the same side first (wick-confirmation) -- a close that clears the level but not that wick doesn't count until a later candle confirms it. A break in the SAME direction as the prevailing bias is tagged BOS (continuation); a break AGAINST the prevailing bias is tagged CHoCH (reversal).
  4. On the 4H chart, note the current bias (bullish or bearish) from the last confirmed 4H break, per steps 2-3.
  5. On the 1H chart, wait for a 1H break that is EITHER (a) a bullish/bearish BOS while the 4H bias already agrees (a continuation entry), OR (b) a bullish/bearish CHoCH while the 4H bias disagrees (a reversal entry) -- AND the 1H candle's close is above its own 50-period EMA for a long, or below it for a short (mandatory on every market, both strategies). Enter at that 1H candle's close.
  6. Place the stop at the low (long) or high (short) of the most recent OPPOSITE-colored 1H candle within the last 30 bars before entry -- i.e. the last candle that would have signaled the other direction. Skip the trade entirely if no such candle exists in that window, or if the resulting stop would already be on the wrong side of entry.
  7. Close the trade the moment the stop from step 5 is hit, OR the 1H chart prints an opposing BOS or CHoCH (against the position's direction) -- whichever happens first.
#3Bitcoin (BTC/USD)Track — Not Yet Tradeable
Variant (LuxAlgo) engine · Strategy A
99 tradesnet avg +0.512Rnet total +50.7Rprob. profit 80.0%risk of ruin 0.1%median DD 21.8%
SummaryHighest RAW score of the three markets downgraded from a live recommendation to "track only" (2026-08-26) -- the Monte Carlo numbers alone look strong (80% probability of profit), but this trades only ~22 times/year with a 21.2% win rate, and its worst historical losing streak (12 trades) would take ~200 real-world days to resolve at that pace. See the Trade Frequency table below for the full comparison across all 5 candidate picks.
StrengthsWider stops make this structurally less fee-sensitive -- the actual mechanism behind why this survives where the original engine doesn't (cost-in-R scales with fee%/stop-width%, not with position size or leverage). The Monte Carlo profile (80% probability of profit, effectively 0% risk of ruin) is genuinely strong ON PAPER -- the problem isn't the edge's shape, it's whether 99 trades over 4.5 years is enough to trust that shape, and whether a trader could actually sit through a ~6.5-month losing stretch to find out.
Weaknesses99 trades is under this project's 200-trade Tier 1 reliability bar -- a full-sample result, not walk-forward validated to gold's standard. Only ~22 trades/year means a realistic bad stretch (12 consecutive losses, already seen once in this history) spans roughly 6.5 months of real calendar time, not weeks -- long enough that a trader has no way to tell, mid-stretch, whether the edge broke or is just doing what a 21.2%-win-rate strategy does sometimes. Crypto markets also carry regime risk (volatility-regime shifts) this backtest window may not fully capture.
ConsiderationsStrategy B has an even larger raw edge (+0.928R/trade, 88.4% probability of profit) on a smaller sample (73 trades, ~16/year) -- same frequency problem, more acute. Leverage does NOT change this market's fee economics (fees and R-risk are both computed per-unit of the underlying, so leverage cancels out of the ratio) -- stop width relative to price is what matters here, not position sizing, and leverage does nothing to fix the frequency problem either. Worth re-evaluating once more history accumulates past the 200-trade bar; not a rejection of the underlying idea, just not practically tradeable at today's sample size and cadence.
Implementation — step by step

Provided for reference, not as a live recommendation -- see Weaknesses/Considerations above for why this combo doesn't clear the tradeability bar yet.

  1. Pull up two charts for Bitcoin (BTC/USD): a 4H chart (for bias) and a 1H chart (for entries/exits) -- both need to stay in sync, so run them on the same feed.
  2. On BOTH charts, track structure breaks (BOS/CHoCH) the same way: Swing highs/lows are confirmed only once price fails to make a new extreme for 50 bars (a rolling fractal-pivot method, matching LuxAlgo's own indicator) -- the pivot's price level is the extreme bar's own high/low, not an adjacent candle's. There is no separate bias variable: each break is classified against whichever direction the PREVIOUS break decided, then immediately becomes the new bias itself.
  3. Every break requires the breaking candle's body to close past the level AND to have closed past the prior candle's own wick on the same side first (wick-confirmation) -- a close that clears the level but not that wick doesn't count until a later candle confirms it. A break in the SAME direction as the prevailing bias is tagged BOS (continuation); a break AGAINST the prevailing bias is tagged CHoCH (reversal).
  4. On the 4H chart, note the current bias (bullish or bearish) from the last confirmed 4H break, per steps 2-3.
  5. On the 1H chart, wait for a 1H break that is EITHER (a) a bullish/bearish BOS while the 4H bias already agrees (a continuation entry), OR (b) a bullish/bearish CHoCH while the 4H bias disagrees (a reversal entry) -- AND the 1H candle's close is above its own 50-period EMA for a long, or below it for a short (mandatory on every market, both strategies). Enter at that 1H candle's close.
  6. Place the stop at the low (long) or high (short) of the most recent OPPOSITE-colored 1H candle within the last 30 bars before entry -- i.e. the last candle that would have signaled the other direction. Skip the trade entirely if no such candle exists in that window, or if the resulting stop would already be on the wrong side of entry.
  7. Close the trade the moment the stop from step 5 is hit, OR the 1H chart prints an opposing BOS or CHoCH (against the position's direction) -- whichever happens first.
#4Euro / US Dollar (EUR/USD)Track — Not Yet Tradeable
Variant (LuxAlgo) engine · Strategy A
85 tradesnet avg +0.107Rnet total +9.1Rprob. profit 58.5%risk of ruin 0.0%median DD 17.6%
SummaryDowngraded from a live recommendation to "track only" (2026-08-26): the original engine has no edge on EUR/USD under either strategy, and the variant engine's different structure definition does flip this specific combo positive -- but only ~19 trades/year, on just 85 total, is too infrequent and too thin a sample to trade responsibly. See the Trade Frequency table below.
StrengthsA genuinely different structure-detection logic succeeding where the original engine fails outright is a real signal, even on a small sample -- not just noise from trying enough variations. Of the three downgraded picks, this one's worst losing streak (7 trades) is the shortest in TRADE-COUNT terms.
WeaknessesThinnest sample of every "clear positive" combo on this page (85 trades, under half the 200-trade bar). 58.5% probability of profit is the weakest Monte Carlo profile among the combos that DO survive costs. At ~19 trades/year, even that comparatively short 7-trade losing streak still spans roughly 4.5 months of real time -- short in trade-count, not in the calendar.
ConsiderationsStrategy B is a reasonable secondary option (61.95% probability of profit, +0.317R/trade) but on an even thinner sample (58 trades, ~13/year) -- same frequency problem, worse. Given the original engine is a confirmed structural loser here, don't be tempted to "fix" it with an entry filter -- the variant engine's different structure logic is the only path that's worked on this market so far, it just hasn't traded enough yet to lean on.
Implementation — step by step

Provided for reference, not as a live recommendation -- see Weaknesses/Considerations above for why this combo doesn't clear the tradeability bar yet.

  1. Pull up two charts for Euro / US Dollar (EUR/USD): a 4H chart (for bias) and a 1H chart (for entries/exits) -- both need to stay in sync, so run them on the same feed.
  2. On BOTH charts, track structure breaks (BOS/CHoCH) the same way: Swing highs/lows are confirmed only once price fails to make a new extreme for 50 bars (a rolling fractal-pivot method, matching LuxAlgo's own indicator) -- the pivot's price level is the extreme bar's own high/low, not an adjacent candle's. There is no separate bias variable: each break is classified against whichever direction the PREVIOUS break decided, then immediately becomes the new bias itself.
  3. Every break requires the breaking candle's body to close past the level AND to have closed past the prior candle's own wick on the same side first (wick-confirmation) -- a close that clears the level but not that wick doesn't count until a later candle confirms it. A break in the SAME direction as the prevailing bias is tagged BOS (continuation); a break AGAINST the prevailing bias is tagged CHoCH (reversal).
  4. On the 4H chart, note the current bias (bullish or bearish) from the last confirmed 4H break, per steps 2-3.
  5. On the 1H chart, wait for a 1H break that is EITHER (a) a bullish/bearish BOS while the 4H bias already agrees (a continuation entry), OR (b) a bullish/bearish CHoCH while the 4H bias disagrees (a reversal entry) -- AND the 1H candle's close is above its own 50-period EMA for a long, or below it for a short (mandatory on every market, both strategies). Enter at that 1H candle's close.
  6. Place the stop at the low (long) or high (short) of the most recent OPPOSITE-colored 1H candle within the last 30 bars before entry -- i.e. the last candle that would have signaled the other direction. Skip the trade entirely if no such candle exists in that window, or if the resulting stop would already be on the wrong side of entry.
  7. Close the trade the moment the stop from step 5 is hit, OR the 1H chart prints an opposing BOS or CHoCH (against the position's direction) -- whichever happens first.
#5Silver (XAG/USD)Track — Not Yet Tradeable
Variant (LuxAlgo) engine · Strategy A
81 tradesnet avg +0.035Rnet total +2.8Rprob. profit 46.4%risk of ruin 0.0%median DD 20.7%
SummaryDowngraded from a live recommendation to "track only" (2026-08-26) -- and the weakest case of the three. The original engine is the single worst combo on the entire site under BOTH strategies (net -234.6R and -398.5R); the variant engine's different structure definition survives costs, but at only ~18 trades/year with a WORSE-than-random-guess 46.4% probability of profit and this page's longest calendar-time losing streak (~358 days, essentially a full year). See the Trade Frequency table below.
StrengthsThe only surviving option for this market at all. Larger sample than the Strategy B alternative (81 vs. 47 trades), giving somewhat more reliable statistics despite a thinner raw edge -- though "more reliable than an even thinner sample" is a low bar.
WeaknessesThin sample (81 trades, well under the 200-trade bar) -- a full-sample result only, not walk-forward validated. Raw edge is thin (+0.035R/trade) and probability of profit (46.4%) sits under 50% even though net avg R is nominally positive. At ~18 trades/year, the historical worst losing streak (18 trades) would take roughly a YEAR of real time to play out -- of the five candidate picks on this page, this is the single hardest one to actually sit through, by a wide margin.
ConsiderationsStrategy B has a materially larger raw edge (+0.548R/trade, 61.0% probability of profit) on an even thinner sample (47 trades, ~10/year) and ranks lower only because the scoring formula weights sample-size confidence -- but it doesn't fix the frequency problem, it makes it worse. This is the weakest-evidenced pick on this page across every dimension (score, sample size, win rate, AND trade frequency); treat as directional only until substantially more history accumulates.
Implementation — step by step

Provided for reference, not as a live recommendation -- see Weaknesses/Considerations above for why this combo doesn't clear the tradeability bar yet.

  1. Pull up two charts for Silver (XAG/USD): a 4H chart (for bias) and a 1H chart (for entries/exits) -- both need to stay in sync, so run them on the same feed.
  2. On BOTH charts, track structure breaks (BOS/CHoCH) the same way: Swing highs/lows are confirmed only once price fails to make a new extreme for 50 bars (a rolling fractal-pivot method, matching LuxAlgo's own indicator) -- the pivot's price level is the extreme bar's own high/low, not an adjacent candle's. There is no separate bias variable: each break is classified against whichever direction the PREVIOUS break decided, then immediately becomes the new bias itself.
  3. Every break requires the breaking candle's body to close past the level AND to have closed past the prior candle's own wick on the same side first (wick-confirmation) -- a close that clears the level but not that wick doesn't count until a later candle confirms it. A break in the SAME direction as the prevailing bias is tagged BOS (continuation); a break AGAINST the prevailing bias is tagged CHoCH (reversal).
  4. On the 4H chart, note the current bias (bullish or bearish) from the last confirmed 4H break, per steps 2-3.
  5. On the 1H chart, wait for a 1H break that is EITHER (a) a bullish/bearish BOS while the 4H bias already agrees (a continuation entry), OR (b) a bullish/bearish CHoCH while the 4H bias disagrees (a reversal entry) -- AND the 1H candle's close is above its own 50-period EMA for a long, or below it for a short (mandatory on every market, both strategies). Enter at that 1H candle's close.
  6. Place the stop at the low (long) or high (short) of the most recent OPPOSITE-colored 1H candle within the last 30 bars before entry -- i.e. the last candle that would have signaled the other direction. Skip the trade entirely if no such candle exists in that window, or if the resulting stop would already be on the wrong side of entry.
  7. Close the trade the moment the stop from step 5 is hit, OR the 1H chart prints an opposing BOS or CHoCH (against the position's direction) -- whichever happens first.
#6S&P 500 (US 500)No Recommendation
No combo currently survives real trading costs
Best gross Strategy Agross avg +0.045Rnet avg -0.042Ravg cost 0.086R
SummaryEvery combo tested on this market -- both structure engines, both strategies -- fails to survive real trading costs. Original engine Strategy A comes closest (gross positive at +0.045R/trade) but Vantage's real S&P 500 index spread (0.086R avg cost) flips it net negative.
StrengthsNone that survive costs -- there is no strength to report for a live-tradeable version of this market today.
WeaknessesAll 4 combos are net negative. It isn't purely a cost problem either: the original engine's Strategy B and both variant-engine strategies show gross weakness before costs are even applied, not just a thin edge overwhelmed by fees.
ConsiderationsThe closest path to viability is a broker with a materially tighter S&P 500 spread than Vantage's Raw ECN -- worth shopping the actual spread before writing this market off entirely. As configured today, though, there is genuinely nothing on this market to recommend running live.