R-value frameworks, execution analysis, and the metrics that separate professional traders from everyone else.
A yen profit shown with a dollar sign is 150 times too big, and a London price in pence read as pounds is 100 times too big. What an international trade needs to record, why R-multiples need no exchange rate, and how SignalDeck now journals stocks on 22 non-US exchanges.
A pasted CSV gives a chatbot a stale snapshot; a connection gives it your actual records. What a connected assistant can read in SignalDeck, what it can change and how it asks first, what it can delete only if you opt in, what it can never do, and three questions to ask it first.
Crypto has no overnight gap, but it has three problems stocks do not: fees taken in the coin that leave phantom open positions, thin hours where the slippage lives, and volatility so different between coins that dollar P&L gives the wrong verdict. Twenty trades where the altcoins made more money and were the worse strategy by four times.
A stop loss on a stock only fires while the market is open, so an earnings gap can turn a planned 1R loss into 2.6R overnight. Twenty trades where three gaps remove 43% of the edge without moving the win rate, the fields no broker export can give you back, and how to tell whether holding overnight pays for your strategy.
The drawdown is the same dollar figure at every stage - what changes, twice, is the moment it gets read. End-of-day on the Test, intraday once you are funded on PRO, end-of-day again on PRO+. The account specs, where the trailing floor locks, what the buffer zone really is, and why a Test pass is evidence gathered under a rule that no longer applies.
The trades a student brings to a review are not a random sample - they are the trades the student remembers, and memorability tracks outcome size, so you see the tails and never the body of the distribution. The four numbers that locate a student, the compliance-partitioned cut that separates a discipline problem from a broken edge, why planned risk has to be tracked separately because the R series cannot see position-size drift at all, and the honesty constraint that makes a public track record worth anything.
Nobody disagrees with "be consistent," which is the first sign the advice is not doing any work. The word does two jobs: process consistency, which no widely used metric measures, and outcome consistency, which is the standard deviation of your R-multiples. SQN scores the second, and SQN is the one-sample t-statistic of your R series - so a tight cluster of losses scores best, the figure drifts upward on sample size alone, and position-size drift is invisible to it by construction.
The twenty-field template deserves to be abandoned, and quitting it was the right call - a field no decision depends on is pure cost. So every field has to pass one test: name the decision it changes, and name the number it feeds across a hundred trades. Five survive. Four make results analysable and a fifth explains the path, the strongest rejected candidate gets a fair hearing, and the honest cost is about ten seconds a trade on top of an import.
A backtest runs on every bar in the sample; your log runs on the few hundred decisions you actually made. One is a dataset, the other an anecdote with timestamps - and the backtest still cannot forecast your account, because it assumes an operator who takes every signal, fills at the modelled price and exits when the rule says so. The four things that happen between the signal and the fill, how to put a number on the gap in R, and the two places our own tooling stops short.
When your log and your broker's records disagree, the statement is right. It is also the answer for every question about money. What it structurally cannot contain is why you entered, what you intended to risk, the trades you passed on, and the rule you broke - none of which ever existed on your broker's side of the connection. Also: why reconstructing them from memory fails in a direction rather than at random, and the one field our own MT4/MT5 sync still cannot pull.
A balance is a sum, and a sum cannot be decomposed after the fact. From our own forward-testing scoreboard: the flagship strategy at 343 trades and -$11,516 with a profit factor of 0.85, next to a 22-trade strategy at +$9,748 and a strategy nobody was watching at +71.4%. Combined, the first two read as a quiet -$1,768 month. Also: why the 22-trade winner is not proven either, the four fields attribution actually needs, and the three cases where the sceptics are right.
Both firms have a "maximum loss" and a "daily loss," and they mean opposite things. Topstep's max loss ratchets up behind your end-of-day close then locks at your starting balance, while its daily limit is only a forced break; The5ers' max loss never moves at all, and on High Stakes the daily limit ends the account. The asymmetry that catches people: Topstep gives unrealised profit no credit but tests the breach on unrealised loss. Worked $50K ratchet table, the side-by-side, and the six journal fields that are genuinely firm-specific.
A +$4,100 month where four transport names made +$12,600 and everything else lost −$8,500 is not a good month, it is one sector call. Why net P&L stops being the headline when the capital is not yours, return on deployed capital separating two traders with identical P&L by more than 3×, the correlated-cluster problem that makes four independently-sized positions one bet, the $2,301 of borrow that quietly eats three quarters of a short's edge, and an honest section on when your firm's own system already does all of it.
Four accounts at +$3,200, +$1,400, −$1,900 and +$800 sum to a healthy-looking +$3,500 month — and hide the evaluation that was already dead at −$1,200. P&L is additive; risk headroom is not. Why edge pools across every account and risk never does, how R makes a $20K evaluation comparable to a $150K funded account, the correlation math that shows four accounts are one bet, the 0.01 lot rounding that silently over-risks your smallest account by 67%, and the readiness test for whether the next account pays for itself.
A "50 pip" stop on gold can mean $50 or $5,000 of real risk depending on whose pip definition your broker uses — a hundredfold range produced by a word. Why gold's contract specs break generic forex journaling, how to size from the contract spec instead of a pip, treating spread as a logged line item rather than a rounding error, a worked example putting a XAUUSD scalp and a EURUSD swing on one R axis, session segmentation around the LBMA auctions, and the prop firm metals rules worth verifying first.
A 2.1 SQN backtest and a 0.6 SQN live run are not automatically two different strategies — first rule out sample variance with a standard-error check, then rank the six real causes (slippage, commissions, look-ahead bias, data quality, regime change, latency), measure your actual slippage, retroactively test for overfitting, separate regime change from a broken edge, and reconcile the two curves.
Your journal says +$1,240 and the dashboard says +$928 — and only one of those decides whether you still have an account tomorrow. The five causes of the gap (commissions, swap, unrealized P&L, the firm's reset clock, partial fills), the difference between balance, equity and a trailing threshold, a ten-minute tie-out that resolves a worked $312 discrepancy, and how to build a headroom number you can trust before entry.
A prop firm's drawdown limit is not a constraint, it is a subsidy — the firm pays to compute your risk in real time and enforce it whether you feel like it or not. Trading your own money, none of that exists. How to derive a self-imposed limit from your own Monte Carlo P10 instead of a round number, the four metrics that replace "did I pass?", and how to keep R-multiples honest across deposits and withdrawals.
MT5, Tradovate, Rithmic, NinjaTrader, cTrader, IBKR, and a dozen regional brokers — every journal claims broad support, and almost none of them tell you what kind of support it actually is. The three connection types (live API, terminal script, CSV) and what each one can and cannot see, an honest platform-by-platform breakdown, and what CSV-only import still gives you.
The most common objection to journaling is that it is unnecessary once you are making money — and it deserves a real answer, not a sales pitch. What net P&L structurally cannot tell you, a ten-minute concentration test you can run on your broker statement today, the four fields that make a log worth keeping, and the three cases where skipping it is genuinely the right call.
Most traders blow an Apex account right after a good session, not a bad one — because the trailing drawdown follows your intraday high-water mark and locks in the peak of a trade you gave back. How the trailing limit works with a worked $50K example, why it ends accounts after winning sessions, and what your journal must track in real time.
Most traders find out whether they were ready by paying the fee and failing. Monte Carlo lets you ask the question first: resample your own trade history across 1,000 simulated attempts, and read your pass probability against the firm's drawdown limit. How to run it, what sample you need, and how to size to survive the rules.
Edgewonk built a loyal following on deep journaling and psychology tagging - but it is now an annual subscription (~$197/yr), its MT4/MT5 sync imports closed trades only, and it has no walk-forward, with a what-if simulator rather than Monte Carlo on your real trades. A fair look at where it still fits and what prop and forex traders are switching to instead.
Most MT5 journals sync through an installed Expert Advisor - closed trades only, no open positions, no live balance while a trade is running. SignalDeck connects via a MetaApi session that sees your account in real time. In a prop challenge, where drawdown is measured on live equity, that connection method is the whole ballgame.
FTMO runs two challenge formats with two different drawdown models. The 2-Step keeps a fixed floor; the 1-Step's floor trails your highest closing balance upward every night and locks in. Confuse the two and you misjudge your headroom at the worst moment. Here's how each reset clock works, with worked $100K examples.
If you keep failing the same challenge, it is almost never the strategy. Five repeatable risk-process failures blow evaluations - unvalidated edge, unknown drawdown headroom, wrong position size, pressure drift, and no pre-trade gate. Here is the fix for each, before you pay the next fee.
Most MT5 prop firm traders are sizing positions against a stale starting balance. SignalDeck connects directly to MT5 -- live balance sync means your Kelly sizing, drawdown headroom, and daily limit display all update as your funded account moves.
TradeZella's backtesting is trade replay, not strategy testing — you place trades bar by bar on historical charts, which measures your discretion rather than the rules. Here is the difference, why it matters when you are trying to establish whether an edge exists, and what real rule-based backtesting looks like.
Most traders who fail an FTMO challenge have a real edge. They fail because they entered trades without knowing their real-time distance to the daily loss limit. This guide covers edge validation, challenge-safe position sizing, live drawdown tracking, and the pre-trade gate that stops violations before they happen.
Your stop gets hit. Price immediately reverses. You were right about the direction but dead anyway. Here's the mechanics behind liquidity hunts, where they form, and how to place stops that survive them — especially on a prop firm challenge.
Maximum Adverse Excursion tells you how far the trade went against you. Maximum Favorable Excursion tells you how much of the move you actually captured. Plot both on a scatter and see the shape of your execution.
Grid search tests every parameter combination in a defined range. Without Walk-Forward Analysis, it reliably finds parameters that look great in-sample and fail live. Here's why — and how to fix it.
Win rate tells you how often you're right. Expectancy tells you how much money you make. A 70% win rate can be a losing strategy. A 35% win rate can be highly profitable. Here's the formula that settles the argument.
Most traders who fail an FTMO Challenge don't fail because their strategy stopped working. They fail because they didn't know — in real time — how close they were to a drawdown limit. Here's what the rules actually say and what your journal must track.
Most forex traders measure results in pips — a unit that changes meaning every time lot size or pair changes. Here's what your forex journal actually needs to capture, why R-multiples are the only unit that aggregates cleanly, and how to convert pips to R on any pair.
R-Multiple turns your stop loss into a unit of measurement. Once every trade is expressed as a multiple of your initial risk, expectancy, SQN, Kelly Criterion, and win rate all start speaking the same language. Here's the full picture.
Run Walk-Forward and Monte Carlo over your own imported trades against a specific firm's rule set to answer the only question that matters before an evaluation: would I pass? Includes a six-tool journal comparison and a verdict by trader profile rather than a single winner.
Profit Factor = Gross Profit / Gross Loss. For every dollar you lose, how many dollars do you make back? Here are the benchmarks, why this single number encodes both your win rate and reward-to-risk ratio, and where profit factor fits alongside SQN and Kelly Criterion.
A backtest optimized on the same data it's judged on isn't evidence — it's a memory test. Walk-forward analysis splits your data into in-sample and out-of-sample windows, then slides forward, so your strategy is always tested on data it was never allowed to see during optimization.
The "never risk more than 2%" rule doesn't know your win rate, your R-ratio, or your current funded balance. Kelly Criterion does — and for prop firm traders, SignalDeck recalculates it automatically as your live MT5 balance moves.
Your Kelly Criterion, Fixed-R sizing, and R-Multiple calculations all start from one number: your current account balance. Log it stale and everything downstream is quietly wrong. Here's what breaks — and how MT4/MT5 real-time sync fixes it.
Backtesting tells you what did happen. Monte Carlo tells you what could happen — including the drawdown your historical sequence happened to avoid. Here's how to stress-test your system before the market runs the experiment for you.
Sizing by dollars feels deliberate. It isn't. Here's the formula that ties every position to a consistent fraction of your account — and why getting this right is the foundation everything else in your journal is built on.
Your strategy worked last quarter. Now it's bleeding. Here are the four data signals that separate a normal drawdown from real edge decay — and the exact protocol for what to do when they appear.
Win rate is the most cited - and most misleading - metric in retail trading. Here's what professional traders measure instead, and why R-multiples change everything.
The SQN score tells you whether your trading system is statistically tradable or just a fluke. Here's how to calculate it from your own trade data, what the benchmarks mean, and how SignalDeck automates it.