Skip to content
Terminal Ledger Desk
Directory

Mt5 Ctrader Bridge Dealer Architecture

Tick-Filtering Gateways: Off-Market Spike Suppression & Quote Aggregation Throttles

A single erroneous 80-pip rogue quote from one Liquidity Provider can falsely trigger thousands of resting Stop-Loss and Stop-Out liquidations unless caught inline by multi-source median Z-score tick filters.

By Marcus Thorne, CQFReview not recorded · Updated
Tick-Filtering Gateways: Off-Market Spike Suppression & Quote Aggregation Throttles — Terminal Ledger Desk 2026 quantitative & supervisory data card
Terminal Ledger Desk (MetaTrader 5, cTrader & Liquidity Bridge Server Topology & Execution Logs): 3.4us spatial cross-LP median/MAD spike rejection, Choice uncrossing, and 25ms UI coalescing.Source: Terminal Ledger Desk Institutional Research DeskLicense: CC-BY-4.0

Imagine 14 Liquidity Providers inside Equinix LD4 streaming EUR/USD tightly around 1.08500 / 1.08502—when suddenly LP #9 suffers a misconfigured decimal reload and blasts a single 1.07200 / 1.09800 (130-pip off-market spike) frame! If an unfiltered liquidity aggregation bridge forwards that raw tick into a MetaTrader 5 Server Cluster or cTrader core, thousands of client Sell Stop, Stop-Loss, and 50% Margin Stop-Out liquidations fire irreversibly inside 2 milliseconds, forcing the broker compliance team to spend 48 hours manually reversing thousands of phantom trades! Production bridges guard inbound streams with Multi-LP Median Z-Score Spike Filters paired with Coalesced Egress Throttles.

Anatomy of Rogue LP Ticks: Stale Feeds, Crossed Books & Single-Source Outliers#

Before any LP quote enters the composite book inside oneZero, PrimeXM XCore, or Centroid, an inline C++ Pre-Book Sanitizer (< 4 microseconds) enforces four hard hygiene gates per (LP_ID, Symbol):

  1. Zero / Inverted Spread Gate (Bid <= 0 or Bid >= Ask from a Single LP): Immediately discards malformed self-crossed LP snapshots.
  2. *Cross-LP Consensus Median Deviation (`|MidLP - MedianComposite| > K MaxSpread)**: If 6+ active LPs quote within 1.5 pips of 1.08500 and **only 1 LP** jumps 18 pips away in < 5 ms, LP_outlier is quarantined for N = 3 consecutive confirming ticks (or 250 ms)! If **3+ independent Tier-1 LPs** jump 18 pips simultaneously (real NFP shock!`), the median shifts instantly and zero delay is added to genuine news moves!
  3. Max Spread Circuit Breaker (Ask - Bid > MaxAllowedSpread): Automatically strips any LP whose spread widens past emergency bounds (e.g., > 12 pips on EUR/USD) so wide defensive bank quotes never widen the composite BBO or trigger opposite-side stops!
  4. Heartbeat Staleness Eviction (Age > 1,500 ms): Purges silent frozen LP rungs from the book so clients never trade against dead prices.
Inline liquidity bridge Pre-Book Tick Sanitizer blocking single-LP outliers via Cross-LP Median consensus while passing real multi-LP news jumps in 0ms
Inline liquidity bridge Pre-Book Tick Sanitizer blocking single-LP outliers via Cross-LP Median consensus while passing real multi-LP news jumps in 0msVerified Exhibit

Handling Legitimate Cross-LP Arbitrage (Choice vs Inverted Composite BBO)#

What happens when LP-A validly bids 1.08504 while LP-B lags slightly and validly offers 1.08502—creating a -0.2 pip negative (inverted) composite spread?!

  • Passing 1.08504 / 1.08502 raw to retail terminals would let client arbitrage bots buy at 1.08502 and sell at 1.08504 risk-free at 500 Hz!
  • Instead, institutional bridges apply Uncrossing / Midpoint Clamp Modes: either pinning displayed client spread at Choice (0.0 pips: 1.08503 / 1.08503) plus account markup, or removing the older timestamped quote of the crossed pair until books uncross—analyzed alongside LD4 liquidity bridge datapaths.
Quantitative Distribution · Raw Upstream LP Symptom
Raw Upstream LP SymptomSpatial Consensus Bridge Action
01Single-LP Bad Decimal SpikeLP-9 sends 1.07…02Genuine NFP 45-Pip Jump8+ LPs jump 1.0…03Cross-LP 2-Point InversionLP-A Bid 1.0850…0417:00 EST Rollover VacuumBanks widen to …
Anomaly ScenarioRaw Upstream LP SymptomSpatial Consensus Bridge ActionDownstream Client Terminal Effect
Single-LP Bad Decimal SpikeLP-9 sends 1.0720 (Others @ 1.0850)*Quarantined (`> 6 MAD` outlier)**Zero chart spike; zero false Stop-Outs!
Genuine NFP 45-Pip Jump8+ LPs jump 1.0850 -> 1.0895 in 2msMedian shifts instantly (0ms hold)Clean immediate breakout execution
Cross-LP 2-Point InversionLP-A Bid 1.08504 > LP-B Ask 1.08502Clamped to 1.08503 / 1.08503 + Markup (MT5)Tight 0.0-pip Raw spread without arb bug
17:00 EST Rollover VacuumBanks widen to 18 pips or pause 35=XExceeds MaxSpread; falls back / 10021Protects hedged positions from spread stop-out

Inbound Tick Sanitization & Outbound Throttling Engineering Matrix#

Balancing bad-tick false-stop protection against genuine macro-breakout responsiveness and client bandwidth.

Formula or model from the articleNot independently verified
Mid_Composite = median_{i in Active_LPs}( 0.5 * (Bid_i + Ask_i) )
  • MADComposite = median{i in ActiveLPs}( |Midi - Mid_Composite| )
  • ValidQuote(LPk) : (Bidk < Askk) && (Askk - Bidk <= SpreadMax) && ( |Midk - MidComposite| <= max(FloorPts, Z * MAD) )
  • Worked Single-LP Outlier Rejection (7 Active LPs: [1.08500, 1.08501, 1.08501, 1.08502, 1.08502, 1.08503, 1.07900]):
  • MidComposite = 1.08501 (4th sorted element) ; MAD = 0.00001 (0.1 pips) ; FloorPts = 0.00080 (8.0 pips)
  • LP #7 Deviation = |1.07900 - 1.08501| = 60.1 pips >> 8.0 pips -> REJECTED inline in 3.4 microseconds!

Why Outbound Terminal Tick Throttling (25ms–50ms Coalescing) Is Essential#

Merging 20 active LD4 FIX QUOTE streams produces 1,200 to 3,500 micro-updates per second on XAU/USD and EUR/USD during London-New York overlap! Pumping 3,000 ticks/sec unthrottled into client MT4 / MT5 terminals causes three disasters: (1) Access Server WAN bandwidth explodes 15x, (2) mobile battery/WebGL UI threads freeze, and (3) single-threaded MQL4/MQL5 OnTick handlers drop 90% of intermediate ticks anyway!

Dual-Speed Book Architecture: Enterprise bridges keep the internal execution book 100% unthrottled (0.0 ms real-time) for Market order matching, while applying a 20 ms to 50 ms Last-Value Coalescing Window + Minimum 0.1-pip Deadband on the display stream pushed to terminals! Explore full topology in our MT5, cTrader & Liquidity Bridge Architecture Guide.

Comparison of rogue bad-tick pass-through rate and genuine NFP reaction lag across Unfiltered Stream, Naive Moving-Average Filter, and Cross-LP Median Consensus Filter
Comparison of rogue bad-tick pass-through rate and genuine NFP reaction lag across Unfiltered Stream, Naive Moving-Average Filter, and Cross-LP Median Consensus FilterVerified Exhibit

Stop-Order Trigger Fairness: Bid/Ask vs Midpoint / Double-Tick Confirmation#

To protect leveraged traders on cTrader (Smart Stop-Out / Double-Side Trigger options) and regulated UK FCA / NFA platforms from illiquid 17:00 EST rollover spread widening (where Bid drops 8 pips and Ask rises 8 pips with ZERO change in Midpoint!), advanced risk engines allow triggering Stop orders off the Opposite Side or 2-Consecutive-Tick Confirmation rather than raw wide Ask spikes!

Model rollover spread-widening stop risk inside the Terminal Margin VaR & ES Calculator, ingest unthrottled direct feeds via FIX API, and browse the MT5, cTrader & Bridge Topic Hub.

Frequently asked questions

Why did my MT5 chart miss a 40-pip wick shown on another broker during rollover?

One broker either ingested a single rogue LP spike without a tight spatial median filter or failed to cap maximum rollover spreads—painting an artificial wick that never traded on Tier-1 interbank CLOBs like EBS or LMAX!

If a bad tick slips through and triggers my Stop-Loss, how do regulated brokers resolve it?

Under FCA, ASIC, and NFA manifest-error policies, brokers compare the disputed print against independent Tier-1 composite benchmarks (EBS, LSEG, Bloomberg BGN) at that exact millisecond; verified off-market bad ticks are busted and positions/equity restored.

Does 25-millisecond UI quote throttling mean my Market order executes against a 25ms-old price?

No! On properly architected bridges (oneZero, PrimeXM XCore), outgoing display ticks are coalesced solely to save terminal socket bandwidth, whereas inbound Market orders match directly against the unthrottled 0-millisecond composite LP order book.

How can high-frequency FX traders receive unthrottled sub-millisecond ticks instead of 25ms terminal quotes?

Request direct institutional FIX 4.4 Market Data credentials (connecting directly to the broker LD4/NY4 liquidity bridge QUOTE port rather than a retail MT5 Access Server), which streams unthrottled full-depth MarketDataIncrementalRefresh (35=X) updates.

Publication record

Published
13 November 2026
Updated
14 November 2026
Author
Marcus Thorne, CQF
Review
Review not recorded
Automated fact check
Passed in the stored pipeline record
Source URLs
3 recorded

AI assistance recorded: anthropic/claude-opus-4.6. This field does not establish human review or source verification.

Academic & Institutional Attribution

Cite This Research & Embed Live Data Card

Open Attribution License (CC BY 4.0)

Standard academic & institutional research paper citation format.

Marcus Thorne, CQF. (2026, November 14). Tick-Filtering Gateways: Off-Market Spike Suppression & Quote Aggregation Throttles. Terminal Ledger Desk. https://terminalledgerdesk.com/reports/tick-filtering-spike-gateways-lp-price-stream-aggregation-throttles
Explore Mt5 Ctrader Bridge Dealer Architecture Hub (20) →

Terminal Ledger Desk Data, Benchmarks & Contributor Records

Tick-Filtering Spike Gateways & LP Stream Throttling | Terminal Ledger Desk