KairosGold

Kairos Gold

Kairos is the ancient Greek word for the opportune moment. Kairos Gold trades XAUUSD (gold) only at statistically selected hours of the day and only in the market regimes where its edge has been verified. The rest of the time it stays flat. It does not trade in unclear conditions.

Live signal

Kairos Gold runs on a real, live account. You can follow or copy every trade on the public signal: https://www.mql5.com/en/signals/2380655

Strategy

Kairos Gold uses ten independent strategy modules. Each module combines an entry hour, a direction, and a market regime, so the system adapts to trending, sideways, and corrective gold markets. Two execution engines are used. The grid engine works on the buy side: it accumulates positions with the same lot size during strong uptrends and closes them with quick profit-taking and basket netting. There is no martingale, and every grid level uses the same lot size. The scalp engine works on both the buy and sell sides with single positions, each protected by a hard stop loss on the server, break-even protection, and a time-based exit. The sell modules harvest pullbacks in bearish regimes, so the system also works in years when gold moves sideways or down. A trend filter based on the daily chart decides which modules are allowed to trade at any given time. When gold offers no clear edge, the system waits.

New in version 1.20: optional all-weather blocks

All ten blocks above are gated by the daily trend filter. That is the intended design, but it has a side effect: stretches appear where no block qualifies. Measured, the regime in which gold sits just above its 200-period daily average (less than +4 ATR away) accounts for roughly a third of all weekdays, and in that band almost none of the ten blocks fire. Over the test period, the longest stretch without a single new position was 52 trading days on the Active preset and 189 trading days, about nine months, on Conservative.

Version 1.20 adds two blocks (11 and 12) that carry no regime condition at all. Block 11 is a clock-driven straddle that places stop orders outside the previous day's high and low; block 12 places stop orders at the break of the current session's range. Both put a stop loss on any position they open and retire it on a schedule. Because they never look at the regime, they work precisely while the ten blocks above are standing down. With them enabled, the longest stretch without a single new position falls from 52 trading days to 13.

These two blocks are off by default. If you are updating from an earlier version, your EA behaves exactly as it did before - verified over the same window and settings with not a single trade different. To use them, load the new All-Weather preset.

New in version 1.30: an on-chart panel, and adjustable trailing for blocks 11 and 12

Version 1.30 adds a small panel in the corner of the chart. It names the current market state and how far gold sits from its 200-period daily average, lists which blocks are trading right now and which are eligible but switched off, and gives the exact price each waiting block is waiting for. It exists because a buyer switched two blocks off as unprofitable after a three-week test in which those two were among only four that were allowed to trade at all - nothing in the terminal showed him that. The panel places no orders and can be hidden.

Blocks 11 and 12 now expose their trailing stop: an on/off switch, and the distance the stop keeps behind price. Switching a trail off does not remove protection - the original stop loss and the scheduled exit both remain. If you set a distance closer to price than your broker accepts for a stop order, the EA places the stop at the broker's minimum and writes one line in the journal, rather than quietly doing nothing.

One honest note about that setting. Thirteen different distances were tested over four and a half years. On the tester's default one-minute model a tighter trail looks steadily better; on real tick data that advantage disappears entirely, and no distance beat the shipped one. The defaults are the only values verified on real ticks.

The preset files have grown from 53 to 58 settings. Please load a preset from the 1.30 pack after updating: MetaTrader does not reset a setting that a preset file does not mention. Updating from 1.20 does not change how your EA trades - compared over the same window and settings, not a single trade differs.

Backtest results

Tested in the MetaTrader 5 tester on the one-minute OHLC model with real spread and commission on an ICM account, XAUUSD, 150,000 JPY starting balance (about 1,000 USD). All six rows below were re-measured on the same build under the same conditions.

Preset | Period | Trades | Win rate | Profit factor | Net gain | Max equity DD

Active | 2025.01-2026.06 | 435 | 93.1% | 4.34 | +135% | 4%

Active | 2022.01-2026.06 | 1,229 | 84.4% | 3.39 | +335% | 13%

Conservative | 2025.01-2026.06 | 258 | 98.8% | 13.25 | +88% | 6%

Conservative | 2022.01-2026.06 | 575 | 91.3% | 5.20 | +148% | 9%

All-Weather | 2025.01-2026.06 | 1,125 | 84.4% | 2.88 | +566% | 16%

All-Weather | 2022.01-2026.06 | 3,369 | 78.0% | 3.09 | +1,661% | 20%

Real-tick verification

The three 2025.01-2026.06 rows above were re-run on the real-tick model (Model 4), the most accurate model the tester offers. Real tick data begins in 2025, so that is the full range for which it exists; the 2022-2026 rows can only be produced on the one-minute model.

Preset | Period | Trades | Win rate | Profit factor | Net gain | Max equity DD

Active | 2025.01-2026.06 real ticks | 408 | 90.9% | 3.60 | +104% | 5%

Conservative | 2025.01-2026.06 real ticks | 250 | 99.2% | 13.84 | +69% | 6%

All-Weather | 2025.01-2026.06 real ticks | 1,097 | 81.3% | 1.97 | +298% | 13%

Real ticks give lower net profit than the one-minute model - most of all for All-Weather, which enters with stop orders. Profit factor holds up better than net profit, and on Conservative it is actually higher on real ticks. If you test this EA yourself, please select "Every tick based on real ticks" and check that History quality in the report reads 100% real ticks. That is the figure worth trusting.

Live and real-tick results will differ from a backtest because of slippage, spread variation and execution. Past performance does not guarantee future results.

An honest note about the All-Weather preset

Over 4.5 years All-Weather returns about five times the net profit of Active, but that gap is not evenly earned. Most of it comes from filling the idle stretches of 2022 and 2023, when Active was almost dormant. In recent conditions the gap is much smaller. Over the six months of 2026.01-2026.06, Active took 62 trades for 58,685 JPY and All-Weather took 146 trades for 68,183 JPY: 2.4 times the trades for 16% more profit, at a profit factor of 1.70 against 5.37.

So the value of All-Weather is not "it always earns more" - it is "it fills the stretches where the ten blocks stand down". When gold is trending, Active is the more efficient choice. All-Weather carries the larger drawdown and the lower profit factor. This is a difference in character, not in quality: Active and Conservative protect quality by declining to trade, while All-Weather exists to cover the waiting. If you can tolerate long flat stretches, use Active or Conservative.

The table above is measured from a 150,000 JPY start, about 1,000 USD. Grid depth is unlocked by the account balance, roughly 800 USD per level up to eight levels, so an account that starts at 6,400 USD or more has the full grid open from the first day and does not travel the same path. Measured over the same 2022.01-2026.06 window from a 1,000,000 JPY start, about 6,700 USD, the Active preset returned +759% with a 22% maximum equity drawdown. If you are starting with a larger account, please read the FAQ entry about it below.

Quick start

Attach Kairos Gold to one XAUUSD chart on any timeframe; the system manages all timeframes internally. Choose a preset: Active for more trades, Conservative for the highest-quality trades only, or All-Weather to cover the waiting stretches. Free preset files are available in the Comments section of this page. Check the journal, where the system prints an environment report and confirms the setup. On the first run the daily history must be loaded, because Kairos Gold uses a daily trend filter and needs about 250 daily bars before it starts trading. If the journal shows that it is waiting for D1 history, open the XAUUSD daily chart once and press Home to scroll back so the broker downloads the history. History is stored per timeframe, so scrolling the H1 chart does not load D1. Once 250 or more daily bars are present, trading begins. This is a one-time step and the most common reason a fresh install does not trade.

Requirements

The symbol is XAUUSD, and the system auto-detects the broker's gold symbol and suffix. A hedging account is required, because netting and exchange accounts are refused. The EA must be attached to a gold chart - this is a requirement, not a recommendation. It can locate the gold symbol from any chart, but break-even, trailing and parts of the exit management are degraded on any other chart, so running it on a non-gold chart is not a supported trading configuration. A starting balance of 1000 USD or more is recommended, or the equivalent in the account currency; it will run from 500 USD, but grid depth scales with balance and trading becomes very sparse - in testing a 500 USD account produced roughly one trade in three months, and 200 USD produced none. Leverage of 1:100 or higher is recommended, ideally 1:500. Broker server time GMT+2 or GMT+3 is recommended, and other time zones are handled by the automatic time adjustment. A VPS is recommended for uninterrupted operation.

Built-in protection

Risk is limited structurally rather than hidden. There is no martingale and no lot multiplication. The grid is scaled to the account balance, so a small account is allowed only a shallow grid, depth unlocks as the account grows, and total exposure is capped. If the bullish regime collapses while grid positions are open, all baskets are closed immediately. New grid positions are not opened into extreme trend extension at the day's high. New entries pause automatically when the spread widens abnormally above its recent norm. On extreme-volatility days, if a single scalp trade's worst-case stop loss would risk more than a set fraction of the account even at the minimum lot size, that trade is skipped. An FOMC news filter pauses new entries from 30 minutes before to 120 minutes after the interest-rate decision, the period when gold can move several dollars within seconds. Open trades are unaffected: stop-loss, take-profit, break-even, the time exit and basket closes all keep working normally inside that window. It is on by default, and an NFP window is available (off by default). A configurable equity stop places a hard floor under the account. A trade-cost watchdog detects within ten trades whether broker commission or spread is consuming the strategy's margin, alerts you, and stops opening new positions. On startup the account type, daily-history depth, symbol specification, and time-zone offset are verified and printed to the journal. The system works on any account currency and converts all internal thresholds automatically.

Every protection above applies unchanged when blocks 11 and 12 are enabled. In addition, an unfilled stop order is treated as an entry that has not executed yet: if the equity stop, the daily loss limit or the cost watchdog fires, any stop orders still resting on the book are cancelled. After an equity stop, these blocks stop placing new orders for 24 hours by default. On Friday they cancel their unfilled orders and close their open positions.

Settings

All strategy logic is built in and pre-tuned. You control which modules are enabled, which the presets do for you, position sizing by Grid base lot per $1000 balance or fixed lot, protection levels such as the equity stop percentage, daily loss limit and maximum grid depth, and trading filters such as maximum spread, the news window, the Friday scalp cut-off and the year-end pause.

From Friday evening the system stops opening new scalp trades. A scalp relies on a time-based exit, and once the market closes for the weekend there is no opportunity to send that exit, so the position would be carried into the following week. Grid baskets are exempt, because closing them as a basket is by design. The cut-off hour is adjustable, and 0 disables it. Only values 1-19 have a practical effect: late-Friday entries from 20:00 onwards are already blocked by a separate built-in rule, so settings of 20-23 add nothing (the startup log notes this). Most users simply load a preset and leave the rest at default.

Honest notes

Kairos Gold performs best in trending gold markets. In flat regimes it deliberately reduces or stops trading, so weeks with few or no trades are normal and intended. The grid engine holds floating positions during pullbacks. Grid legs carry no individual stop-loss; they are closed as a basket by the EA itself (quick take-profit, basket netting, regime exit, equity stop), so the EA has to stay running - a VPS is recommended. The equity stop and exposure scaling limit this risk, but grid trading is never risk-free. Please leave position sizing on automatic (fixed lot at 0) and never set the equity-stop percentage to 0. Please test on a demo account first and start with the Conservative preset.

One more thing about blocks 11 and 12. When they earn, they raise the account balance, and grid depth is unlocked by balance - so they indirectly deepen the grid. Measured over the same window, grid add-on legs rose from 186 to 469. That is a designed interaction rather than a fault, but it is part of why All-Weather carries a larger drawdown than Active.

Frequently asked questions

Which preset does the live signal use? The public signal runs the Active preset (all ten modules, with blocks 11 and 12 off).

Does updating to version 1.30 change how my EA behaves? No. The on-chart panel is display only - it places no orders - and the two new trailing settings ship at exactly the values the previous version used internally. Before and after the update were compared over the same window and settings, on both the regime blocks and the straddle blocks, with not a single trade different. Please still reload your preset after updating, because presets saved under 1.20 do not contain the five newer settings.

Does updating to version 1.20 change how my EA behaves? No. Blocks 11 and 12 are off by default, and all four existing presets state them as off explicitly. Before and after the update were compared over the same window and settings with not a single trade different. Only load the All-Weather preset if you want the new blocks. Please reload your preset after updating in any case, because presets saved earlier do not contain the new settings.

Who is the All-Weather preset for? Anyone who cannot tolerate long stretches with no trades. You gain trades and active days, and you give up drawdown and profit factor. Please also read the honest note above: in recent conditions the profit gap to Active has narrowed considerably. Run it on a demo for a few weeks first and check that the higher trade frequency and the floating-loss levels suit you.

The EA is not trading, why? It needs about 250 daily bars before its trend filter starts. If the journal shows it waiting for D1 history, open the XAUUSD Daily chart once and press Home. Quiet stretches with few or no trades are also normal by design, because it only trades at its selected hours and in the regimes where it has an edge. If it is the quiet stretches themselves you want to shorten, consider the All-Weather preset.

It is only taking sell trades, or only buys, is that a fault? No. A daily-chart trend filter decides which modules may trade: in a strong uptrend the buy grid and buy scalps are active, when price sits below its daily trend the sell scalps take over, and in unclear conditions it waits. So an all-sell or all-buy stretch simply reflects the current regime.

Which broker and account do I need? Any broker offering XAUUSD, on a hedging account. Raw or low-spread accounts match the backtest cost basis best; a wider-spread Standard account also works, but net profit runs somewhat lower.

Can I run it with less risk, or without the grid? Yes. The grid is only Strategy block 1 and block 2. Turn them off and it runs single-position scalp trades only, each with a hard stop. To hold fewer positions at once, enable fewer blocks. Removing the grid lowers the drawdown but also the return, so it is a trade-off.

I am starting with 6,400 USD or more, should I change anything? Yes, one setting. Grid depth is unlocked by the balance and is fully open from about 6,400 USD, so a large account begins at full exposure instead of growing into it, and at full depth the normal floating loss comes close to the default 30% equity stop. Set Grid base lot per $1000 balance to 0.005 instead of the default 0.01. Measured over 2022.01-2026.06 that lowered the maximum equity drawdown from 22% to 19%, and in the worst single year measured, 2024, from 34.3% to 20.0%, turning that year from -12.9% into +15.5%; about 70% of the profit remains. Lower the lot rather than the equity stop: for a grid that carries no stop-loss on the individual legs, the equity stop is the last floor under the account.

Which preset for steady, controlled growth? Conservative, which is the most selective, with the highest profit factor and the gentler drawdown. It also waits the longest: over the test period its longest stretch without a single new position was 189 trading days.

Does it work on a cent account? Yes. Position sizing is derived from the broker's tick value, so it adapts to a cent account automatically and the parameters stay the same. Please verify the lot sizes look sensible on a demo first.

Why does it stop trading around a rate decision? That is the news filter added in version 1.10. It pauses only new entries from 30 minutes before to 120 minutes after the FOMC announcement; trades already open keep being managed normally. The decision dates are built in and verified through 2027, after which the filter simply stops applying rather than guessing. If you are updating from an earlier version, please reload your preset - preset files saved before 1.10 do not contain the new settings.

More from this developer

Expert Advisors:

Aurum Quartet, XAUUSD gold: https://www.mql5.com/en/market/product/187293

BullionForge, XAUUSD gold: https://www.mql5.com/en/market/product/182722

WeekAnchor Eight, eight JPY and CHF crosses: https://www.mql5.com/en/market/product/180321

NightAnchor, USDCHF and other night straddles across twenty pairs: https://www.mql5.com/en/market/product/185922

YenAnchor GBPJPY, free: https://www.mql5.com/en/market/product/182789

PhaseFlow Fusion, XAUUSD and BTCUSD: https://www.mql5.com/en/market/product/183202

Live signals:

Kairos Gold: https://www.mql5.com/en/signals/2380655

BullionForge: https://www.mql5.com/en/signals/2379035

Aurum Quartet: https://www.mql5.com/en/signals/2386955

NightAnchor: https://www.mql5.com/en/signals/2380654

WeekAnchor: https://www.mql5.com/en/signals/2378669

Seller page: https://www.mql5.com/en/users/masa999/seller

If anything is unclear, please leave a message in the Comments; I answer personally. I develop and support this alone from Japan, so a reply can take a day or two across time zones, but every question gets an answer. Honest feedback of any kind is truly welcome. Whether your review is positive or critical, it helps me improve the product and is genuinely appreciated. Thank you for your interest in Kairos Gold.