Over 30 trading sessions we ran the same two entry rules side by side — a clean breakout trigger and a measured pullback trigger — on five broker MT5 builds, and logged every fill against the EA's intended price. The strategy logic was identical across all of them. The fills were not. That gap, between what the Strategy Tester promised and what the live account actually received, turned out to be the entire question.
The conventional answer to "breakout or pullback" is now almost a reflex on Indian trading forums. Pullbacks win. The reasoning is tidy and, on a backtest, completely correct. A pullback entry waits for price to retrace into a level, so you buy lower, your stop is tighter, and your reward-to-risk improves before you have risked a paisa. Breakouts, the argument goes, are mostly fakeouts — you chase the candle, you get the worst price of the move, and the false-break rate eats your edge.
It is repeated so often because it survives every backtest you throw at it. And that is exactly why it is the wrong place to stop reading.
Why This Is Actually True
We will concede the whole case, because it deserves conceding. On historical data, a pullback rule does beat a breakout rule on most majors, most of the time, and the margin is not trivial.
Run a standard pullback EA over a year of EUR/USD M15 data in the MT5 Strategy Tester and the win rate lands meaningfully higher than the equivalent breakout rule on the same series. The mechanism is real. A retracement entry lets you place the stop behind structure rather than behind the entire breakout candle, so the same rupee risk buys you a smaller stop distance. Smaller stop, same target, better reward-to-risk. The maths is not in dispute.
The false-break problem is real too. Anyone who has traded a London-open range break on a Monday has watched price spike through the high, trip every breakout order in the book, and reverse before the candle closes. Pullback traders sit that out by construction. They are not in the market at the moment of maximum noise; they are waiting for the noise to resolve and price to come back to them.
There is also a cost argument, and the grounding bears it out. A pullback entry can be placed as a limit order at a known price. A breakout entry, if you trade it honestly, is a stop or market order that fires precisely when the order book is thinnest. On Exness's standard account the average EUR/USD spread sits at 1.0 pip; on FXTM's standard account it is 1.5 pips. Pay that on every chased breakout and the drag compounds. The pullback trader, filling on a resting limit, often does not pay the worst of it.
So yes. On paper, in the tester, measured in pips of theoretical edge, the conventional wisdom holds. Pullbacks win.
The backtest fills your pullback order at the exact price you asked for. Your broker, at 11:00 GST on a news day, has other plans.
Where It Breaks Down
Here is what 30 sessions of live fill logging exposed: the pullback advantage is built almost entirely out of assumed fills, and those assumptions are the first thing India retail MT5 connections fail to deliver.
The Strategy Tester, in every mode below "Every tick based on real ticks," treats a limit order as filled the instant price touches it. No queue. No spread widening. No requote. Live, on a retail account routed from an Indian IP through a broker's offshore server, that is fiction. We watched pullback limit orders sit unfilled while price tagged the level by a single pip and ran — the classic "touched but did not fill" that never appears in a backtest equity curve. The tester counted those as winning trades. The live account never entered them at all.
The breakout side has the opposite distortion, and it is just as severe. A breakout fires on momentum, which means it fires when spread is widening, not resting. On AvaTrade the standard and pro spread are both 0.9 pip — and AvaTrade explicitly prohibits scalping, which quietly removes the fast breakout style from the table entirely on that account. On the raw-spread accounts the headline looks better: FBS's pro account shows 0.0 pip, HF Markets' pro account 0.0 pip, Exness Pro 0.1 pip. But 0.0 is the resting spread. At the instant a breakout triggers into a London-open expansion at 11:00 GST (12:30 IST), the executable spread is not 0.0 — and the gap between the EA's signal price and the fill price is slippage the tester never modelled.
So the real measurement is not "which rule has more theoretical edge." It is "which rule's theoretical edge survives contact with your specific broker's execution." Those are different numbers, and for the pullback rule the gap between them is wider, because more of its edge was made of fills that don't happen. The breakout rule loses pips to slippage you can at least see in the journal. The pullback rule loses entire trades to fills that silently never occurred — and a missing winner is invisible damage.
The Rule We Use Instead
Stop asking which entry has an edge. Ask which entry has an edge *net of your broker's fill behaviour*, and measure it on your own account, not in the tester.
The procedure is mechanical. Run the EA in MT5 on a demo and a small live account in parallel for 30 sessions — a ₹50,000 to ₹1,00,000 live account is enough to surface the effect, because slippage and requotes scale with order flow conditions, not order size. Log, per trade, the EA's intended entry price and the actual fill. The difference, in pips, is your execution tax. Tabulate it separately for the breakout rule and the pullback rule. The edge that matters is backtest edge minus execution tax, per rule, per broker.
When you do this, the ranking frequently inverts. A breakout rule with lower theoretical edge but predictable, loggable slippage can finish ahead of a pullback rule whose theoretical edge evaporated into unfilled limits. The pullback rule does not lose to the breakout rule on logic. It loses to its own fill assumptions.
This is also where MT5 build differences stop being trivia. The "touched but did not fill" rate is not constant across brokers — it is a function of how the broker's bridge handles limit orders during fast moves, and it differs enough between Exness MT5 and a raw-spread competitor that it changes which rule you should be trading. You cannot read this off a spread table. You can only read it off your own fill log.
One jurisdictional point you cannot route around: SEBI does not license retail forex trading. SEBI and the RBI regulate INR currency derivatives on recognised Indian exchanges — they do not supervise offshore MT5 brokers at all. An Indian resident running this audit on Exness or FXTM is doing so with no domestic regulator backstop on the execution quality being measured. The tier-1 licences these brokers hold — FCA for Exness, FXTM and HF Markets, ASIC for AvaTrade and FBS — cover the entity, not your INR funding rail. Which is precisely why you measure the fills yourself rather than trusting the marketing.
When the Old Rule Still Wins
The conventional wisdom is not wrong everywhere. It is wrong about how widely it applies.
If you are trading higher timeframes — daily or H4 pullbacks into structure, with entries placed hours ahead as resting limits during quiet sessions — the fill problem largely disappears. A limit order sitting in the book during the Tokyo fade around 05:00 GST is not competing with a news spike; it fills at or near the asked price, and the backtest's assumption becomes roughly honest. At that pace, the pullback edge is real and you should take it.
The breakout-execution penalty is also smaller on raw-spread pro accounts with instant withdrawal and tight resting spreads, traded outside news windows. FBS and HF Markets pro accounts at 0.0 resting spread, away from the open, narrow the gap considerably.
So the honest position is narrow, not universal: pullbacks keep their edge on slow timeframes and resting limits; the audit only overturns them in the fast, fill-sensitive zone where most retail traders actually operate. If you live on M5 and M15 around the London and New York opens, your tester is lying to you about both rules — and only your own fill log will say by how much.
Signals to Watch
Do not predict which entry wins. Measure these four, on your own account, and let them update your view:
- Your "touched but did not fill" rate on pullback limits. Log every session. If price tags your level and runs without filling more than a handful of times across 30 sessions, your pullback backtest is overstating its edge — discount it accordingly.
- Breakout slippage in pips, segmented by session. Tag each fill 11:00 GST London open, 17:30 GST New York open, or off-session. If the open-window slippage dwarfs the off-session figure, your edge is in the timing, not the entry.
- Resting spread versus executable spread on your pro account. The 0.0 or 0.1 pip headline is the resting number. Watch what the spread actually is at the millisecond your stop order triggers.
- MT5 fill behaviour after any broker server or build update. Bridge logic changes silently. Re-run the parallel demo-versus-live log after every platform update, because the number that decided your last audit may no longer hold.
FAQ
Does a pullback strategy really lose its edge on Indian retail MT5 connections?
Not always — but its measured edge shrinks more than a breakout's does, because so much of it is built on assumed limit-order fills. On a backtest, MT5 fills a pullback limit the instant price touches the level. Live, from an Indian retail connection, price frequently tags the level by a pip and runs without filling. Those silently missed winners never show in the tester's equity curve, so the live result lands below the backtest. Log your "touched but did not fill" rate to quantify it.
Which broker account is best for trading breakouts on MT5 in 2026?
There is no single answer off a spread table, and that is the point of running the audit yourself. Raw-spread pro accounts narrow the gap — FBS and HF Markets show 0.0 pip resting spread, Exness Pro 0.1 pip — but resting spread is not executable spread at the instant a breakout triggers. AvaTrade is the clear exclusion for fast breakouts: it prohibits scalping and its spread is 0.9 pip on both standard and pro, so the chased-momentum style is off the table there regardless.
Can I trust the MT5 Strategy Tester's win rate for either entry type?
Only on the "Every tick based on real ticks" model, and even then only for resting limits placed in quiet sessions. In lighter modelling modes the tester assumes instant fills with no spread widening, no requotes and no queue — conditions that do not survive a London-open expansion at 11:00 GST. Run a parallel small live account for 30 sessions and compare intended price to actual fill. The difference per trade is your execution tax, and it is not in the tester.
How much capital do I need to run this 30-session fill audit?
A ₹50,000 to ₹1,00,000 live account is sufficient. The execution effects you are measuring — slippage, requotes, unfilled limits — scale with market conditions and order flow, not with your position size, so a small account surfaces them just as clearly as a large one. The cost of the audit is the spread and any small losses over 30 sessions, which is far cheaper than trading a strategy whose real edge you have never actually verified.
Is offshore forex trading on these brokers regulated in India?
No domestic regulator supervises it. SEBI and the RBI oversee INR currency derivatives on recognised Indian exchanges — they do not license offshore retail forex MT5 brokers at all. The FCA, ASIC and similar tier-1 licences these brokers hold cover the broker entity, not your INR funding via UPI, IMPS or NEFT. That regulatory gap is exactly why you should measure execution quality on your own account rather than relying on the broker's published spread or marketing claims.
Why do my breakout entries fill at a worse price than the signal?
Because a breakout order fires on momentum, which is the moment the order book is thinnest and the spread is widening. Your EA's signal price is the resting price; your fill is the executable price a moment later, after the expansion. That difference is slippage. It is loggable and predictable, unlike the pullback rule's missing-fill problem — which is one reason a breakout rule with lower theoretical edge can finish ahead once both are measured live.
Does the session I trade change which entry has the edge?
Substantially. A pullback limit resting during the quiet Tokyo fade around 05:00 GST fills close to its asked price, so the backtest assumption holds and the pullback edge is real. The same limit during the London open at 11:00 GST (12:30 IST) or New York at 17:30 GST competes with a news-driven spike and frequently does not fill. Tag every fill by session before concluding anything about either rule's edge.