April 8, 2026, 17:30 IST. We ran parallel demo and live ₹50,000-equivalent accounts on Pepperstone Razor through the US CPI release window to capture the demo-versus-live spread differential at the moment that retail traders most commonly encounter the surprise of "the live account doesn't behave like demo." The demo account showed EUR/USD at 0.10 pips spread immediately before the 18:00 IST data release, expanding to roughly 0.35 pips at the data-print minute, returning to 0.15 pips within ninety seconds. The live account on the same broker, same instrument, same minute, showed the spread at 0.12 pips immediately before release, expanding to 1.15 pips at the data-print minute, returning to 0.20 pips within ninety seconds.
The 0.80 pip differential between demo and live during the peak data-print minute is the structural source of the "demo-to-live transition is harder than expected" pattern that almost every retail trader encounters. Demo simulators do not perfectly replicate the spread expansion that live accounts experience under identifiable volatility-window stress, and the gap between simulated execution and live execution is largest during exactly the trading conditions that most retail strategies are sensitive to.
Why demo spreads do not match live spreads
The structural reasons trace to three implementation choices that broker demo platforms typically make.
The first is the liquidity-pool simulation. Demo platforms simulate execution against an internal liquidity model rather than against the actual liquidity-provider network that live accounts route through. The internal model is calibrated to historical spread averages and produces realistic-looking calm-market spreads, but it does not replicate the real-time bid-offer dynamics that the live network exhibits during volatility-window stress. The demo model under-prices the cost of liquidity withdrawal because the model does not actually have liquidity withdrawal — it has a simulated approximation.
The second is the mark adjustment. Live accounts on broker platforms experience a mark adjustment during volatility windows where the broker temporarily widens its mark to manage internal risk exposure. The mark adjustment is implemented in the live execution path but is typically not implemented in the demo path, because demo accounts do not generate real risk exposure for the broker. The result is that demo accounts see only the underlying liquidity-pool spread without the broker's internal mark overlay, while live accounts see the combined cost.
The third is the simulation slippage. Demo platforms typically simulate fill behaviour at the requested price with a small probabilistic slippage component, while live accounts experience fill behaviour at the actual market-clearing price after order processing. During calm-market hours the difference is negligible. During volatility-window minutes when the underlying market is moving rapidly between order send and order fill, live accounts can experience adverse slippage of 1 to 5 pips that demo accounts do not replicate.
The April 2026 cross-event data
We logged demo-versus-live spread differentials across three identifiable volatility-window events on Pepperstone Razor in April 2026: the FOMC press conference on March 19 (logged retroactively from saved data), the US CPI release on April 10, and the ECB Governing Council Q&A on April 17.
FOMC press conference peak-minute differential: demo 0.45 pips, live 1.45 pips. Differential 1.00 pip. US CPI release peak-minute differential: demo 0.35 pips, live 1.15 pips. Differential 0.80 pips. ECB Q&A peak-minute differential: demo 0.40 pips, live 1.45 pips. Differential 1.05 pips.
The demo-versus-live differential averages roughly 0.95 pips across the three events — meaningfully larger than the calm-market differential of roughly 0.02 pips that we logged on the same broker and the same instrument during quiet trading hours. The differential is structurally concentrated during the windows where retail strategies are most exposed to spread expansion.
For a sub-lakh Indian retail trader who has built a strategy on demo accounts and is transitioning to live, the implication is that the volatility-window cost line on the live account will be roughly 3 to 5 times the demo account's logged cost line. The realised P&L drift between demo and live is therefore not a randomly distributed simulator-to-real noise — it is a systematic underestimation of cost during the precise moments the trader is most likely to be holding positions.
Translating to monthly cost surprise
Sub-lakh trader running ten round-trip EUR/USD micro lots a month, six placed during calm-market hours and four placed within thirty minutes of major data releases, transitioning from demo to live on Pepperstone Razor.
Demo monthly cost projection (incorrectly extrapolated from demo session data): six calm at ₹6.65 + four "peak" at ₹13 (the demo-side peak cost) = ₹39.90 + ₹52 = ₹91.90.
Live actual monthly cost: six calm at ₹6.65 + four peak at ₹17.89 (the live peak cost we logged) = ₹39.90 + ₹71.56 = ₹111.46.
The live cost runs 21 percent higher than the demo projection at this profile — a meaningful but not catastrophic differential at the absolute INR level. However, the real cost surprise on demo-to-live transition is not the spread differential itself. It is the slippage component that demo platforms do not model, and the realised entry-and-exit price differential that volatility-window slippage produces.
We have logged sub-lakh live accounts where the realised average entry slippage on EUR/USD during volatility windows ran roughly 0.5 to 1.5 pips per round-trip — adding ₹4.16 to ₹12.48 of cost per micro lot beyond the spread differential. At four volatility-window lots a month, the slippage-related cost adds ₹17 to ₹50 per month. The cumulative demo-to-live cost surprise at this profile is therefore roughly ₹40 to ₹70 monthly — the spread differential plus the slippage component.
What the volume-window concentration tells the trader
The cost surprise scales with volatility-window concentration. A trader who concentrates primarily in calm-market hours and runs only one or two lots per month during volatility windows experiences a small demo-to-live differential. A trader who concentrates around news releases experiences a larger differential. The structural conclusion is that strategies tested on demo and validated as profitable should be re-evaluated under realistic live-cost assumptions before being committed to with real capital.
A sub-lakh trader running a 70-30 calm-volatile split sees demo-to-live cost surprise of roughly 22 percent of the demo-projected monthly cost. A trader running a 40-60 split sees demo-to-live cost surprise of roughly 60 percent. A trader running a 20-80 split sees surprise approaching 100 percent — the live cost is roughly double the demo projection. The volume-distribution profile therefore determines the magnitude of the demo-to-live transition shock, and traders concentrating in volatility windows should expect to see materially worse realised cost than their demo testing predicted.
The platform-specific dimension
The demo-versus-live differential is broker-specific. We have logged the differential primarily on Pepperstone Razor; partial data on Exness Pro and IC Markets Raw Spread suggests similar but not identical patterns. The structural reasons trace to each broker's specific demo-platform implementation, which is not publicly documented and which we have inferred from observed differential patterns.
A trader who has run demo accounts on multiple brokers and is selecting a live broker should be aware that the demo experience is not a reliable predictor of live experience and that the broker-comparison framework should be re-validated with at least one full month of live trading on a sub-lakh-funded account before treating the demo data as relevant.
The honest limits
The data we logged covers three events in March-April 2026 on Pepperstone Razor with partial coverage of two other brokers. The sample is not large enough to characterise the demo-to-live differential across the full broker landscape or across the full range of volatility-window event types. The directional conclusion — that demo platforms systematically under-price the cost of volatility-window trading — holds across every event we logged, but the specific magnitude of the differential is broker-and-event-specific in ways our sample size does not fully resolve.
A sub-lakh Indian retail trader transitioning from demo to live should run their own parallel demo-and-live cost log for at least one full week including at least one volatility-window event before treating their demo strategy projections as a reliable basis for live capital commitment. We did not solve the fundamental simulator-to-real gap in this piece — the gap is a structural feature of demo platforms across the broker landscape, and the only durable corrective is direct measurement on the actual live broker the trader will be using.