the final test of any trading system: forward performance.
a strategy that works in back-testing is necessary but not sufficient. a strategy that works in the last three months of live trading is also not sufficient. the real test is forward performance over a full market cycle.
a full market cycle includes: a trending period, a ranging period, a high-volatility period, a low-volatility period, and at least one unexpected event (flash crash, policy surprise, data shock).
why back-testing overstates performance:
— you selected the strategy because it worked on this data. this is data-mining bias.
— you know where the major events were, so you may have (consciously or not) avoided the periods that would have been hardest.
— transaction costs in live trading are worse than assumed.
the only resolution: forward testing with real money (at small size) over a sufficiently long period. 100 trades is a minimum sample. 200+ gives more statistical confidence.
the traders who believe their strategy before 200 live trades are almost always working from insufficient evidence. the traders who have 500+ live trades documented with a clear process have actual information about their edge.
a strategy that works in back-testing is necessary but not sufficient. a strategy that works in the last three months of live trading is also not sufficient. the real test is forward performance over a full market cycle.
a full market cycle includes: a trending period, a ranging period, a high-volatility period, a low-volatility period, and at least one unexpected event (flash crash, policy surprise, data shock).
why back-testing overstates performance:
— you selected the strategy because it worked on this data. this is data-mining bias.
— you know where the major events were, so you may have (consciously or not) avoided the periods that would have been hardest.
— transaction costs in live trading are worse than assumed.
the only resolution: forward testing with real money (at small size) over a sufficiently long period. 100 trades is a minimum sample. 200+ gives more statistical confidence.
the traders who believe their strategy before 200 live trades are almost always working from insufficient evidence. the traders who have 500+ live trades documented with a clear process have actual information about their edge.
what central bank credibility means and why it matters for FX.
central bank credibility is the market's belief that the bank will do what it says it will do.
a credible central bank: when it signals a rate hike, rates actually rise. when it commits to a target, it defends the target. inflation expectations in a credible bank's currency remain anchored.
an incredible central bank: markets discount its forward guidance. rate decisions regularly surprise because the bank's signals are unreliable. inflation expectations drift because the market doesn't trust the bank will maintain its mandate.
in FX: a currency backed by a credible central bank tends to have lower volatility and more predictable reaction functions. the BOJ's credibility problem in 2022-2024 (extreme dovishness despite high inflation) contributed to JPY weakness because the market correctly discounted the BOJ's ability to normalize.
why this matters: when a central bank makes a commitment (YCC level, inflation target, floor), you need to assess whether they will actually defend it. history of credibility, political independence, and balance sheet capacity are the inputs. the SNB in 2011 was credible right up until January 2015 — a reminder that credibility has limits under sufficient pressure.
central bank credibility is the market's belief that the bank will do what it says it will do.
a credible central bank: when it signals a rate hike, rates actually rise. when it commits to a target, it defends the target. inflation expectations in a credible bank's currency remain anchored.
an incredible central bank: markets discount its forward guidance. rate decisions regularly surprise because the bank's signals are unreliable. inflation expectations drift because the market doesn't trust the bank will maintain its mandate.
in FX: a currency backed by a credible central bank tends to have lower volatility and more predictable reaction functions. the BOJ's credibility problem in 2022-2024 (extreme dovishness despite high inflation) contributed to JPY weakness because the market correctly discounted the BOJ's ability to normalize.
why this matters: when a central bank makes a commitment (YCC level, inflation target, floor), you need to assess whether they will actually defend it. history of credibility, political independence, and balance sheet capacity are the inputs. the SNB in 2011 was credible right up until January 2015 — a reminder that credibility has limits under sufficient pressure.
📌 The "Trap" strategy: trading on major news events
Application: Non-Farm Payrolls (NFP) — the most significant news event, producing strong volatility.
Execution:
▪️ Two to three minutes before the release, place two limit orders (buy and sell) near the current price, with tight stop-losses on both.
▪️ After the data is published, monitor the market reaction.
▪️ On a sharp directional move, act immediately (within a fraction of a second):
— cancel the opposite order;
— move the stop-loss to break-even.
▪️ Then monitor the move for the first 5–10 minutes (maximum). Take profit manually once the target scenario develops.
Key risks (must be considered):
▪️ Slippage. At the moment of publication, limit orders may fill at a price significantly different from the one set.
▪️ Spread widening. In the first few seconds the spread widens sharply, which can trigger stop-losses before a stable move has formed.
▪️ Broker execution. Requotes, delays, or temporary blocking of order execution are possible during high volatility.
▪️ Position sizing. Given the above risks, trading a minimal position size is recommended.
Key condition:
This strategy is executed manually only. Automation and delayed decision-making are not possible — the setup plays out within seconds.
Recommended only for traders with a high level of experience and fast reactions. Not recommended for beginners; practise on a demo account first.
⚠️ This material is for informational purposes only and does not constitute personal financial advice. Trading on financial markets carries a high risk of capital loss.
Application: Non-Farm Payrolls (NFP) — the most significant news event, producing strong volatility.
Execution:
▪️ Two to three minutes before the release, place two limit orders (buy and sell) near the current price, with tight stop-losses on both.
▪️ After the data is published, monitor the market reaction.
▪️ On a sharp directional move, act immediately (within a fraction of a second):
— cancel the opposite order;
— move the stop-loss to break-even.
▪️ Then monitor the move for the first 5–10 minutes (maximum). Take profit manually once the target scenario develops.
Key risks (must be considered):
▪️ Slippage. At the moment of publication, limit orders may fill at a price significantly different from the one set.
▪️ Spread widening. In the first few seconds the spread widens sharply, which can trigger stop-losses before a stable move has formed.
▪️ Broker execution. Requotes, delays, or temporary blocking of order execution are possible during high volatility.
▪️ Position sizing. Given the above risks, trading a minimal position size is recommended.
Key condition:
This strategy is executed manually only. Automation and delayed decision-making are not possible — the setup plays out within seconds.
Recommended only for traders with a high level of experience and fast reactions. Not recommended for beginners; practise on a demo account first.
⚠️ This material is for informational purposes only and does not constitute personal financial advice. Trading on financial markets carries a high risk of capital loss.
the cost of switching strategies during a drawdown.
one of the most reliably destructive patterns in retail trading: a strategy underperforms for 4-6 weeks. the trader decides the strategy is broken and switches to a new one.
the new strategy performs well for 3 weeks. then it underperforms.
repeat.
the pattern means the trader is always entering strategies at local highs and exiting at local lows. they capture the losing periods of each strategy and miss the recovering periods.
the math is brutal: strategy A has a 3-month drawdown, then recovers to new highs. the trader exits at the bottom of the drawdown and enters strategy B at the top of its recent run. they then experience strategy B's drawdown and exit that too.
the solution is not to be loyal to a bad strategy. it is to have a defined evaluation period and criteria before you start — and to hold to them. "i will evaluate this strategy after 100 trades" is a plan. "i will stop if it doesn't work in 2 weeks" is not a plan — it is emotional risk management masquerading as discipline.
one of the most reliably destructive patterns in retail trading: a strategy underperforms for 4-6 weeks. the trader decides the strategy is broken and switches to a new one.
the new strategy performs well for 3 weeks. then it underperforms.
repeat.
the pattern means the trader is always entering strategies at local highs and exiting at local lows. they capture the losing periods of each strategy and miss the recovering periods.
the math is brutal: strategy A has a 3-month drawdown, then recovers to new highs. the trader exits at the bottom of the drawdown and enters strategy B at the top of its recent run. they then experience strategy B's drawdown and exit that too.
the solution is not to be loyal to a bad strategy. it is to have a defined evaluation period and criteria before you start — and to hold to them. "i will evaluate this strategy after 100 trades" is a plan. "i will stop if it doesn't work in 2 weeks" is not a plan — it is emotional risk management masquerading as discipline.
FOMC minutes: why they're read even though the market barely reacts to them.
the minutes are a detailed account of the fed meeting, published at 14:00 US eastern time (ET). there's usually no strong reaction on release: unlike the payrolls report, the minutes tend not to produce sharp candles.
the value of the minutes lies not in the immediate reaction but in the signal about future direction.
what gets taken from them: — the distribution of votes. the rate decision is made by a vote; each committee member's position shows the balance of opinion inside the fed. — the policy direction. the text helps to gauge the trend going forward across several markets: currencies, the dollar index (DXY), bonds, gold.
reaction sequence: the minutes hit the bond market first. gold reacts not to the document itself but to the subsequent moves in the dollar index and treasury yields. a fall in treasury yields is the condition for a corresponding move in gold. 📊
context at the time of release: the latest non-farm payrolls report — 57k versus the expected 110k.
educational content, not investment advice. trading carries a risk of loss.
the minutes are a detailed account of the fed meeting, published at 14:00 US eastern time (ET). there's usually no strong reaction on release: unlike the payrolls report, the minutes tend not to produce sharp candles.
the value of the minutes lies not in the immediate reaction but in the signal about future direction.
what gets taken from them: — the distribution of votes. the rate decision is made by a vote; each committee member's position shows the balance of opinion inside the fed. — the policy direction. the text helps to gauge the trend going forward across several markets: currencies, the dollar index (DXY), bonds, gold.
reaction sequence: the minutes hit the bond market first. gold reacts not to the document itself but to the subsequent moves in the dollar index and treasury yields. a fall in treasury yields is the condition for a corresponding move in gold. 📊
context at the time of release: the latest non-farm payrolls report — 57k versus the expected 110k.
educational content, not investment advice. trading carries a risk of loss.
on humility.
the market has been in operation, in some form, for centuries. the participants include the most well-resourced institutions in the world, with teams of economists, quants, and traders who have spent decades refining their process.
within this context, confidence that you have identified something these participants have missed is usually a warning sign, not an opportunity.
this is not an argument for passivity. it is an argument for calibration.
where there are consistent, small edges for disciplined retail traders:
— process discipline (institutions are not perfect; their flows can be anticipated around known events)
— non-consensus positioning (retail consensus often runs counter to eventual market direction)
— timeframe advantages (institutional flows operate over different holding periods; short-term noise can be traded)
where there are not edges for retail:
— competing on speed
— having better macro data
— accessing order flow
the traders who survive long-term are those who have an accurate map of where their edge is and is not. the ones who don't survive have an inflated view of the edge they carry relative to the market they're trading in.
the market has been in operation, in some form, for centuries. the participants include the most well-resourced institutions in the world, with teams of economists, quants, and traders who have spent decades refining their process.
within this context, confidence that you have identified something these participants have missed is usually a warning sign, not an opportunity.
this is not an argument for passivity. it is an argument for calibration.
where there are consistent, small edges for disciplined retail traders:
— process discipline (institutions are not perfect; their flows can be anticipated around known events)
— non-consensus positioning (retail consensus often runs counter to eventual market direction)
— timeframe advantages (institutional flows operate over different holding periods; short-term noise can be traded)
where there are not edges for retail:
— competing on speed
— having better macro data
— accessing order flow
the traders who survive long-term are those who have an accurate map of where their edge is and is not. the ones who don't survive have an inflated view of the edge they carry relative to the market they're trading in.
the trade you took that you shouldn't have.
there is usually a moment you can identify, looking back, where you knew the trade wasn't quite right but took it anyway.
the thesis was incomplete. the setup was missing one confirmation you normally require. or the risk/reward was marginal at best.
but you took it because:
— you hadn't traded in a while and needed to "do something"
— you'd been watching the pair for hours and felt entitled to a move
— you didn't want to miss the opportunity
— you'd already sized it in your head and mentally committed
this category of trade — the one you knew was substandard — tends to lose more often than your normal trades. not because the market is punishing you, but because your normal filters exist for good reason. when you bypass them, you are taking lower-quality setups.
the record is useful here: if you tag your trades with "full setup" vs "borderline," the data will almost certainly show that borderline trades underperform. this is the empirical argument for enforcing your own criteria, not the moral one.
there is usually a moment you can identify, looking back, where you knew the trade wasn't quite right but took it anyway.
the thesis was incomplete. the setup was missing one confirmation you normally require. or the risk/reward was marginal at best.
but you took it because:
— you hadn't traded in a while and needed to "do something"
— you'd been watching the pair for hours and felt entitled to a move
— you didn't want to miss the opportunity
— you'd already sized it in your head and mentally committed
this category of trade — the one you knew was substandard — tends to lose more often than your normal trades. not because the market is punishing you, but because your normal filters exist for good reason. when you bypass them, you are taking lower-quality setups.
the record is useful here: if you tag your trades with "full setup" vs "borderline," the data will almost certainly show that borderline trades underperform. this is the empirical argument for enforcing your own criteria, not the moral one.
building a trading plan: the minimum viable document.
a trading plan should be short enough to actually use and specific enough to prevent discretionary drift under pressure. here is the minimum structure:
— strategy definition. what is the setup? one paragraph that describes it precisely. not vaguely ("I look for good setups") but specifically ("I look for pullbacks to the 50 EMA on the 4h chart in pairs that are trending on the daily, with the entry triggered by a reversal candle and RSI divergence").
— universe. which pairs do you trade? why these pairs specifically?
— risk per trade. the dollar amount, not the percentage. "I risk $150 per trade."
— stop placement. how do you determine the stop? ATR-based? below the recent swing low?
— target logic. how do you set the target? 2× risk? prior structure? trailing stop?
— session. which hours do you trade? what are the off-limits times (major releases, thin sessions)?
— review cadence. when do you review your trades? daily? weekly?
this document should exist in written form. it should be consulted before each trade session. when you deviate from it, the deviation should be documented and reviewed. without this document, you are not trading a plan — you are improvising in real time under financial pressure.
a trading plan should be short enough to actually use and specific enough to prevent discretionary drift under pressure. here is the minimum structure:
— strategy definition. what is the setup? one paragraph that describes it precisely. not vaguely ("I look for good setups") but specifically ("I look for pullbacks to the 50 EMA on the 4h chart in pairs that are trending on the daily, with the entry triggered by a reversal candle and RSI divergence").
— universe. which pairs do you trade? why these pairs specifically?
— risk per trade. the dollar amount, not the percentage. "I risk $150 per trade."
— stop placement. how do you determine the stop? ATR-based? below the recent swing low?
— target logic. how do you set the target? 2× risk? prior structure? trailing stop?
— session. which hours do you trade? what are the off-limits times (major releases, thin sessions)?
— review cadence. when do you review your trades? daily? weekly?
this document should exist in written form. it should be consulted before each trade session. when you deviate from it, the deviation should be documented and reviewed. without this document, you are not trading a plan — you are improvising in real time under financial pressure.
gold and real yields: the basic mechanism.
gold pays no coupon, so its opportunity cost is the real yield on government bonds — the rate after inflation. when real rates rise, holding gold gets more expensive: capital shifts into instruments that pay. a stronger dollar reinforces the effect — the metal is priced in dollars, and its dollar price falls.
the flip side is protective and structural demand. in periods of uncertainty, gold absorbs flows out of risk assets, and structural demand from large holders acts as a counterweight.
how this ties to macro data: weak employment data lowers rate expectations → the dollar and real yields stop climbing → opportunity-cost pressure eases. a reversal in real rates and the dollar brings the previous pressure back.
the transmission channel remains the real rates ↔️ dollar link; it works particularly hard during tightening cycles, when the rise in real yields outpaces demand for gold as an inflation hedge.
gold pays no coupon, so its opportunity cost is the real yield on government bonds — the rate after inflation. when real rates rise, holding gold gets more expensive: capital shifts into instruments that pay. a stronger dollar reinforces the effect — the metal is priced in dollars, and its dollar price falls.
the flip side is protective and structural demand. in periods of uncertainty, gold absorbs flows out of risk assets, and structural demand from large holders acts as a counterweight.
how this ties to macro data: weak employment data lowers rate expectations → the dollar and real yields stop climbing → opportunity-cost pressure eases. a reversal in real rates and the dollar brings the previous pressure back.
the transmission channel remains the real rates ↔️ dollar link; it works particularly hard during tightening cycles, when the rise in real yields outpaces demand for gold as an inflation hedge.
on patience.
there are two types of patience in trading. one makes money. the other costs it.
the first: patience in waiting for the right setup. declining to trade when conditions don't meet your criteria. watching the market, doing nothing, and being comfortable with that. this is productive patience. it conserves capital for genuine opportunities.
the second: patience in holding a bad trade. waiting for a position that is clearly wrong to "come back." adding to it. refusing to take the loss. this is passive avoidance masquerading as patience. the cost compounds with every session the position sits.
distinguishing them:
— the first involves waiting before the trade, with no current exposure.
— the second involves waiting during the trade, with existing exposure and an evolving thesis.
when you're waiting on an existing position, ask: "if I had no position, would I enter this trade right now at current price?" if the answer is no, the patience is the wrong kind. the position should be exited.
there are two types of patience in trading. one makes money. the other costs it.
the first: patience in waiting for the right setup. declining to trade when conditions don't meet your criteria. watching the market, doing nothing, and being comfortable with that. this is productive patience. it conserves capital for genuine opportunities.
the second: patience in holding a bad trade. waiting for a position that is clearly wrong to "come back." adding to it. refusing to take the loss. this is passive avoidance masquerading as patience. the cost compounds with every session the position sits.
distinguishing them:
— the first involves waiting before the trade, with no current exposure.
— the second involves waiting during the trade, with existing exposure and an evolving thesis.
when you're waiting on an existing position, ask: "if I had no position, would I enter this trade right now at current price?" if the answer is no, the patience is the wrong kind. the position should be exited.
the difference between a signal and noise.
in a market that generates thousands of data points per day, the ability to distinguish signal from noise is the core analytical skill.
signal: information that genuinely updates your probability estimate of where price is going. a central bank communication shift. a structural break in positioning. a macro data print that changes the rate path pricing.
noise: normal, statistically expected variation in price that has no predictive value for the next move. a 15-pip move in the absence of news. an economic reading that merely confirms the existing consensus.
the problem: noise looks like signal in real time. both produce chart patterns, both trigger alerts, both can be rationalized with a story.
the filter:
— does this information change anything about the underlying drivers of the pair? if not, it is noise.
— is this move larger than 1 ATR in less than 1 session without a news catalyst? possibly significant. with news: expected.
— is there a reason why this level or move would be observed in a market where no one has any information? if yes, it's likely noise.
professional traders develop calibrated skepticism about what constitutes a genuine signal. most things in markets are not.
in a market that generates thousands of data points per day, the ability to distinguish signal from noise is the core analytical skill.
signal: information that genuinely updates your probability estimate of where price is going. a central bank communication shift. a structural break in positioning. a macro data print that changes the rate path pricing.
noise: normal, statistically expected variation in price that has no predictive value for the next move. a 15-pip move in the absence of news. an economic reading that merely confirms the existing consensus.
the problem: noise looks like signal in real time. both produce chart patterns, both trigger alerts, both can be rationalized with a story.
the filter:
— does this information change anything about the underlying drivers of the pair? if not, it is noise.
— is this move larger than 1 ATR in less than 1 session without a news catalyst? possibly significant. with news: expected.
— is there a reason why this level or move would be observed in a market where no one has any information? if yes, it's likely noise.
professional traders develop calibrated skepticism about what constitutes a genuine signal. most things in markets are not.
Important this week:
🇺🇸 The "official" US corporate earnings season gets underway
Monday 13 July:
🇪🇺 Adoption of the EU's 21st sanctions package against Russia is expected
🇪🇺\🇺🇦 Meeting of the Coalition of the Willing on Ukraine
🛢 OPEC Monthly Oil Market Report – 14:00
Tuesday 14 July:
🇺🇸 US – CPI inflation (June) – 15:30
🇺🇸 Warsh to speak before the House Financial Services Committee – 17:00
Wednesday 15 July:
🇺🇸 US – PPI inflation (June) – 15:30
🇺🇸 Warsh to speak before the Senate – 17:00
Thursday 16 July:
🇺🇸 US – Retail Sales (June) – 15:30
🇺🇸 US – Pending Home Sales (June) – 17:00
Friday 17 July:
🇺🇸 US – Housing Starts (June) – 15:30
✴️🇺🇸 Hearings on the CLARITY Act, the US crypto market structure bill, are expected in the US Congress
🇺🇸 The "official" US corporate earnings season gets underway
Monday 13 July:
🇪🇺 Adoption of the EU's 21st sanctions package against Russia is expected
🇪🇺\🇺🇦 Meeting of the Coalition of the Willing on Ukraine
🛢 OPEC Monthly Oil Market Report – 14:00
Tuesday 14 July:
🇺🇸 US – CPI inflation (June) – 15:30
🇺🇸 Warsh to speak before the House Financial Services Committee – 17:00
Wednesday 15 July:
🇺🇸 US – PPI inflation (June) – 15:30
🇺🇸 Warsh to speak before the Senate – 17:00
Thursday 16 July:
🇺🇸 US – Retail Sales (June) – 15:30
🇺🇸 US – Pending Home Sales (June) – 17:00
Friday 17 July:
🇺🇸 US – Housing Starts (June) – 15:30
✴️🇺🇸 Hearings on the CLARITY Act, the US crypto market structure bill, are expected in the US Congress
what winning in FX actually looks like over a year.
let's be precise.
a professional retail FX trader, trading 100-150 setups per year with a 55% win rate and an average risk/reward of 1.5:1, generates approximately 30-40% gross return on deployed capital before friction costs.
friction costs (spread, slippage, overnight carry in some positions) typically consume 5-15% of gross return depending on strategy type and volume.
net result: 20-30% annual return on deployed capital for a disciplined, consistently-applied strategy.
this is exceptional performance by any financial standard. it is not the 200% in 3 months that is typical in retail trading content. it is not the "signals group" result of 500 pips this week.
the reason realistic numbers matter: if your expectation is 200% in year one, a 25% return will feel like failure. you will change strategies, overtrade, or take excessive risk to close the gap. calibrated expectations are not pessimistic — they are the precondition for making decisions that lead to the actual returns.
let's be precise.
a professional retail FX trader, trading 100-150 setups per year with a 55% win rate and an average risk/reward of 1.5:1, generates approximately 30-40% gross return on deployed capital before friction costs.
friction costs (spread, slippage, overnight carry in some positions) typically consume 5-15% of gross return depending on strategy type and volume.
net result: 20-30% annual return on deployed capital for a disciplined, consistently-applied strategy.
this is exceptional performance by any financial standard. it is not the 200% in 3 months that is typical in retail trading content. it is not the "signals group" result of 500 pips this week.
the reason realistic numbers matter: if your expectation is 200% in year one, a 25% return will feel like failure. you will change strategies, overtrade, or take excessive risk to close the gap. calibrated expectations are not pessimistic — they are the precondition for making decisions that lead to the actual returns.