Clients crossing the 50% loss threshold
Per 1,000 initial clients · lower is better
Fewer severe losses. More capital preserved.
Both results hold in all 18 variants. At baseline pricing, severe losses fall from an average of 371 to 2 per 1,000 clients. Average final capital rises from $37,914 to $42,096.
Complement traditional leverage with an offering that preserves more client capital in the simulated comparison. CRL maintains 1× exposure before the trigger and amplifies the result to 5× at the trigger. Fewer severe losses are the measured outcome; translating that into longer client relationships is the opportunity still to be tested. Margins and risks are analysed separately for A-book, B-book and hybrid models.
Both results hold in all 18 variants. At baseline pricing, severe losses fall from an average of 371 to 2 per 1,000 clients. Average final capital rises from $37,914 to $42,096. In the simulated strategy, CRL positions are closed when the trigger is reached. The results reflect this specific trading strategy, not every possible use of CRL.
Per 1,000 initial clients · lower is better
Dollars per initial client · higher is better · starting capital $50,000
Comparison conditions and notes →
The cost of leverage and spread →
Cumulative margin per initially acquired client, after costs included in the model and before acquisition costs, taxes and the cost of capital. Filters in the detailed sections do not change this overview.
| Book | Traditional | CRL | How to interpret it |
|---|---|---|---|
| A-book | 1,555 $ | 2,127 $ |
Higher margin in the model
An increase of $572; favourable in 18/18 variants. The executable hedging price still needs to be verified. |
| Pure B-book | 11,434 $ | 6,542 $ |
Profitable on average, but below traditional at baseline pricing
CRL produces positive margin in 15/18 variants; on average $4,892 less. At Q + 5%: about 95% of traditional margin and an average of 16 versus 371 severe losses per 1,000 clients. |
| Hybrid · 10% B-book | 2,543 $ | 2,569 $ |
Small average advantage
Only $26 more; favourable in 7/18 variants. The 10% share is an illustrative scenario, identical for both products. |
In the simulated comparison, CRL beats traditional 5× leverage on severe losses and client capital in all 18 variants at baseline pricing. A-book earns more at the model coverage price; in B-book, the broker chooses the price and the trade-off.
Charts, distributions and all 18 variants at baseline pricing.
18 correlated pairs, 36 distinct client scenarios, 36,000 simulated 24-month paths and 864,000 client-month observations. Three historical windows; one random seed. These are simulations, not independently observed clients.
Client pricing: commission of $20 per million per side, CRL service fee of 2.75 bp, annual funding of 3% of notional, no additional spread. Provider priced at theoretical Q. Voluntary churn disabled: trading activity and commercial retention are separate measures.
Counts per 1,000 initial clients; averages across variants may be fractional. −50% is relative to the initial $50,000, not a drawdown from the peak. In-position capital depletion is assigned to the closing month; balances reflect closed trades, not mark-to-market. Monthly anniversaries run from 2 January at 08:00 UTC.
STD / CRL columns always follow the same order. Monthly results are distinct from cumulative results.
The table of 18 variants remains complete even when a single cohort is selected in the charts. Averages of medians and percentiles describe the cohorts, not the percentile of an independent pooled sample.
The broker earns service revenue and transfers the amplified payment to a provider under the model assumptions. Coverage is assumed to be purchased at opening: Q theoretically prices a EUR/USD one-touch, while the modelled contract requires the incremental 4× payment to be calculated at the actual execution price after the trigger, including any barrier overshoot. Hedging only the client's 1× exposure does not fund this payment. The price and availability of this specific coverage remain to be verified with the provider.
24 months · 1,000 clients per cohort · average of 18 correlated variants · 5× leverage and baseline pricing. External share: provider at theoretical Q and execution of 5 seconds + 0.5 pip. Unquoted costs, taxes and cost of capital excluded. Calculator selections do not change this section.
Positive CRL margin in 18/18 variants; above traditional in 18/18 variants. CRL margin range: $1,725 / $2,449 per client. These are not confidence intervals.
In the simulated comparison, the client-capital benefit is accompanied by higher broker margin than traditional leverage in all 18 variants under the stated conditions. CRL A-book margin comes from the 2.75 bp service fee and commissions, including the exit commission calculated on 5× notional, net of included costs. This provides a concrete basis for comparing the economic budget with broker and provider terms.
The client already pays Q; the additional budget is the headroom for a higher coverage cost before reaching parity with traditional leverage. Under the presented stress (5 seconds + 0.5 pip), the average advantage reaches zero at approximately 0.60% above Q: thresholds across variants are 0.35–0.89% of Q, not percentage points of capital. Under baseline execution, the average threshold is approximately 1.9% above Q (1.45–2.27% across variants). These thresholds hold client pricing, trades and other costs fixed; they indicate parity with traditional leverage, not zero CRL margin. Different client pricing requires a separate assessment.
The book model changes the broker's accounting, not the product or the client paths. Average capital at 24 months: CRL $42,096, traditional $37,914, from an initial $50,000. Fewer severe losses do not mean fewer people making a loss; commercial retention still needs to be verified. Full client KPIs →
Dollars per initial client, after only the included costs. Same scale across all three sections. Solid blue: CRL; dashed grey: traditional.
| Month | Traditional monthly $ | CRL monthly $ | Traditional cumulative $ | CRL cumulative $ | Cumulative difference $ |
|---|---|---|---|---|---|
| 1 | 79 | 108 | 79 | 108 | 29 |
| 2 | 62 | 84 | 141 | 191 | 50 |
| 3 | 67 | 117 | 208 | 308 | 101 |
| 4 | 56 | 71 | 264 | 379 | 115 |
| 5 | 58 | 129 | 322 | 508 | 186 |
| 6 | 63 | 105 | 385 | 613 | 228 |
| 7 | 61 | 101 | 447 | 715 | 268 |
| 8 | 70 | 118 | 516 | 833 | 317 |
| 9 | 63 | 67 | 580 | 901 | 321 |
| 10 | 50 | 62 | 629 | 963 | 333 |
| 11 | 56 | 68 | 686 | 1,031 | 345 |
| 12 | 38 | 64 | 723 | 1,095 | 372 |
| 13 | 64 | 106 | 787 | 1,201 | 414 |
| 14 | 67 | 68 | 854 | 1,269 | 415 |
| 15 | 93 | 98 | 948 | 1,366 | 419 |
| 16 | 72 | 71 | 1,020 | 1,437 | 418 |
| 17 | 79 | 101 | 1,098 | 1,538 | 440 |
| 18 | 73 | 118 | 1,171 | 1,657 | 486 |
| 19 | 72 | 79 | 1,243 | 1,735 | 492 |
| 20 | 68 | 110 | 1,311 | 1,846 | 535 |
| 21 | 73 | 68 | 1,384 | 1,913 | 529 |
| 22 | 66 | 65 | 1,450 | 1,979 | 529 |
| 23 | 65 | 68 | 1,515 | 2,046 | 531 |
| 24 | 40 | 81 | 1,555 | 2,127 | 572 |
Transferring the payoff does not eliminate execution and liquidity risks. In the sample, both the worst daily loss and the maximum cumulative decline are higher with CRL. Provider reimbursement delays may require cash advances.
| Measure based on realised cash flows | Traditional $ | CRL $ |
|---|---|---|
| Worst loss over a complete day | 4,167 | 91,236 |
| Maximum cumulative decline from peak | 4,167 | 102,212 |
| Maximum initial cash deficit | 0 | 10,927 |
Extremes across the 18 variants, not quantiles or capital requirements. Cash flows recognised at closes and modelled due dates: open exposures are not continuously revalued. Maxima may occur on different dates and in different variants and must not be added together. Definitions, trigger concentration and stress →
To be agreed with the broker and provider: coverage price and exact payout, trigger reference (broker quotes or an independent source), provider creditworthiness and collateral, slippage and reimbursement timing. Cost of capital, acquisition and support remain to be verified. Select A-book in the shared calculator.
The broker retains the full counterparty result. Premium, funding and payoff are reconciled with client wealth without double counting; no external hedge is purchased.
24 months · 1,000 clients per cohort · average of 18 correlated variants · 5× leverage and baseline pricing. No external hedge or hedge execution cost is charged to the pure B-book. Unquoted costs, taxes and cost of capital excluded. Calculator selections do not change this section.
Positive CRL margin in 15/18 variants; above traditional in 0/18 variants. CRL margin range: -$283 / $12,467 per client. These are not confidence intervals.
CRL generates a positive margin in 15 of the 18 variants, but less than traditional in all 18. As a complementary offering it can contribute margin; for migrated clients, the traditional margin forgone must also be accounted for.
A positive margin does not demonstrate an advantage over traditional. The premium, exit and replacement tests did not identify a joint average advantage in client wealth and broker margin under the conditions examined.
The book model changes the broker's accounting, not the product or the client paths. Average capital at 24 months: CRL $42,096, traditional $37,914, from an initial $50,000. Fewer severe losses do not mean fewer people making a loss; commercial retention still needs to be verified. Full client KPIs →
Dollars per initial client, after only the included costs. Same scale across all three sections. Solid blue: CRL; dashed grey: traditional.
| Month | Traditional monthly $ | CRL monthly $ | Traditional cumulative $ | CRL cumulative $ | Cumulative difference $ |
|---|---|---|---|---|---|
| 1 | 1,157 | -265 | 1,157 | -265 | -1,422 |
| 2 | 981 | -117 | 2,138 | -382 | -2,520 |
| 3 | -531 | 107 | 1,607 | -276 | -1,883 |
| 4 | 1,032 | 924 | 2,639 | 649 | -1,990 |
| 5 | 612 | 281 | 3,250 | 930 | -2,320 |
| 6 | -30 | 494 | 3,220 | 1,424 | -1,796 |
| 7 | 46 | 223 | 3,266 | 1,647 | -1,619 |
| 8 | 1,411 | 396 | 4,677 | 2,043 | -2,634 |
| 9 | 1,232 | 290 | 5,909 | 2,333 | -3,576 |
| 10 | 588 | 30 | 6,498 | 2,363 | -4,135 |
| 11 | 50 | 299 | 6,547 | 2,662 | -3,886 |
| 12 | 948 | -12 | 7,495 | 2,650 | -4,846 |
| 13 | -48 | 20 | 7,448 | 2,669 | -4,778 |
| 14 | 579 | 761 | 8,027 | 3,430 | -4,597 |
| 15 | -182 | 1,272 | 7,845 | 4,702 | -3,143 |
| 16 | 958 | 1,193 | 8,803 | 5,895 | -2,908 |
| 17 | 116 | 548 | 8,919 | 6,442 | -2,477 |
| 18 | 27 | -845 | 8,946 | 5,597 | -3,349 |
| 19 | 637 | 507 | 9,583 | 6,104 | -3,479 |
| 20 | 109 | -562 | 9,692 | 5,542 | -4,150 |
| 21 | 290 | 339 | 9,982 | 5,881 | -4,101 |
| 22 | 504 | -36 | 10,486 | 5,845 | -4,641 |
| 23 | 41 | 21 | 10,526 | 5,865 | -4,661 |
| 24 | 908 | 676 | 11,434 | 6,542 | -4,892 |
CRL reduces the worst simulated daily loss, but increases the maximum cumulative decline from peak and the initial cash deficit. Preserving client capital does not automatically reduce every aspect of counterparty risk.
| Measure based on realised cash flows | Traditional $ | CRL $ |
|---|---|---|
| Worst loss over a complete day | 1,411,410 | 621,961 |
| Maximum cumulative decline from peak | 4,333,381 | 5,513,276 |
| Maximum initial cash deficit | 794,816 | 3,675,386 |
Extremes across the 18 variants, not quantiles or capital requirements. Cash flows recognised at closes and modelled due dates: open exposures are not continuously revalued. Maxima may occur on different dates and in different variants and must not be added together. Definitions, trigger concentration and stress →
Risk capital, exposure limits, liquidity, commercial costs and the share of new versus migrated clients. Select pure B-book in the shared calculator.
Additional assumptions. These results do not replace those in the overview.
No configuration simultaneously improves average client wealth, relationship duration and positive B-book margin relative to traditional under the conditions examined. Replacement creates acquisition costs, but also new margins: both are included. The original premium remains the baseline. This result does not establish impossibility for every product or broker model.
With the original premium, exit after a realised loss of 50% and replacement from the following month, CAC would have to exceed $16,771 for CRL margin to exceed traditional. However, CRL margin becomes negative above $6,534: there is no viable interval. The initial client retains $3,561 more on average. Exit thresholds are scenarios, not observed churn.
The same product, two income statements compared. The selector starts at the original premium: client and broker always use the same configuration. Other pricing levels are explicit sensitivities and do not replace the A-book / B-book / hybrid baseline.
Initial CAC of 0: a control without paid acquisition, not a commercial estimate. It is identical for both products and also applies to initial clients. The delay is a scenario and does not depend on future results.
The same people, the same 24 months. Includes withdrawn capital.
Initial and replacement clients, per initial client slot, over the full period.
CRL: solid blue · Traditional: dashed grey.
The broker can decline an uneconomic replacement. A separate check optimises this choice at the original premium. Even with this option, no CAC interval offers all three benefits together. The calculator here shows systematic replacement; the full optional-replacement check is included in the technical dossier.
Counts per 1,000 initial client slots; margins in dollars per initial slot. Wealth: average for initial clients only, including withdrawals. Replacement clients are different people making new deposits; these are not top-ups by the same clients.
Totals across all clients, per initial slot. Aggregate wealth is not a return comparison: the number of people and the duration differ.
17 pricing levels × 6 exit rules × 5 replacement policies = 510 configurations, each calculated across the 18 correlated historical variants: 9,180 comparisons and 220,320 paired monthly observations. These are economic projections derived from existing ledgers, not new independent market simulations.
Key assumption: subsequently acquired clients follow the same average curves for their configuration by relationship age. Results are calculated separately for each variant and then averaged. This is not a backtest with new entries on actual historical dates. No terminal value and no arbitrary retention revenue are included.
Internal checks: reconstruction without replacement, directly summed cohorts, capacity limit, cash-flow conservation including CAC and source hashes. The independent review of the baseline does not cover this extension. Taxes and the cost of risk capital remain excluded.
External context, not calibration: Plus500, FY2025 preliminary results, p. 2 reports AUAC of $1,267. This is a group average, not the marginal CAC of a CRL client depositing $50,000. It is not used as a central assumption.
Actual relationship duration is distinct from capital preserved. This analysis extends only the B-book; A-book and hybrid models retain the conditions disclosed in their respective sections.
At the technical threshold of $500, nobody exits within 24 months: no lifetime extension emerges relative to traditional. In the explicit scenario of exit after a realised loss of 50%, with premium Q + 7.5%, restricted mean duration over 24 months is 23.93 months versus 21.55; 970.6 clients per 1,000 remain versus 647.9. B-book margin is $12,762 versus $10,846 per initial client. Longer duration and higher margin coexist in 17/18 variants. This result is conditional on the exit rule, not an estimate of actual churn. Average client wealth, including withdrawals, is $35,963 versus $38,536: longer duration does not imply higher final capital.
Exits here are loss-tolerance scenarios: the same threshold applies to both products. After the trade that takes the balance below the threshold, the client withdraws the remaining capital and opens no further positions. No new deposits or replacement clients are introduced.
Duration is restricted to the 24 observed months; clients who do not exit contribute 24 months. This is neither full average lifetime nor a commercial-retention forecast. Thresholds use the post-close balance, not the minimum during a position.
Client wealth = balance still in the account + capital already withdrawn. Counts and amounts averaged across the 18 correlated variants. Clients trading during the month can outnumber those remaining at month-end. Margin after operating/licence costs and a fixed cost of $100 over 24 months per initial client, retained even after exit; before acquisition, taxes and cost of capital.
Initial value zero: no benefit added. V is the average value over the same horizon that is not already included in the ledger. It must subtract the net margin forgone on replacement clients from the replacement cost avoided. CAC alone is not V. This is an economic threshold, not a replacement simulation or an LTV estimate. V does not change charts, balances or exits.
17 admissible pricing levels × 18 pairs × 6 exit rules. Previously verified ledgers are truncated at the first exit event; before that event, every trade and decision remains identical. No new calibration of leverage, stops, trigger, market or behaviour.
Sources document relationships between results, activity and cessation, but do not calibrate our exit thresholds or their prevalence among clients. Source CB01 and the documentary audit distinguish relative hazard, cessation, withdrawals and ruin. CB03 concerns another population and is not used to impose a Forex curve. No empirical churn rate has been invented.
Verified: no-exit replay, return of remaining capital, counts, censored durations and the monthly identity: client wealth + B-book margin + included costs = initial capital. The independent review covers the baseline; this extension has internal checks. The package retains all 18 variants, all pricing levels and every month.
Outcome: The economic assessment conditional on exit rules is complete. An extension of actual commercial lifetime has not been demonstrated; exit frequencies and net replacement value for the broker's population are needed to turn the sensitivity into a forecast.
90% follows A-book accounting and 10% remains internalised. The share is identical for CRL and traditional: there is no selection of better clients, no routing optimisation and no added netting benefit.
24 months · 1,000 clients per cohort · average of 18 correlated variants · 5× leverage and baseline pricing. External share: provider at theoretical Q and execution of 5 seconds + 0.5 pip. Unquoted costs, taxes and cost of capital excluded. Calculator selections do not change this section.
Positive CRL margin in 18/18 variants; above traditional in 7/18 variants. CRL margin range: $1,845 / $3,291 per client. These are not confidence intervals.
Average CRL margin is slightly above traditional. The advantage appears in 7 of 18 variants: a positive average does not imply a consistent advantage.
The 10% share is an illustrative scenario, not an estimated optimum. The 90% externally covered share depends on the same coverage availability and pricing assumption as A-book; however, the hybrid budget must be assessed against its own margin, not the 0.60% threshold of pure A-book. Additional costs can absorb the small average advantage. Other shares can be explored in the economic calculator with trades held fixed, without resimulating clients. With internalised share fixed, loading_max = Δhybrid / [(1 - internalised share) × Q] per variant; average parity is about 0.03% above Q. Some variants are already below traditional at price Q. This threshold concerns parity, not zero CRL margin.
The book model changes the broker's accounting, not the product or the client paths. Average capital at 24 months: CRL $42,096, traditional $37,914, from an initial $50,000. Fewer severe losses do not mean fewer people making a loss; commercial retention still needs to be verified. Full client KPIs →
Dollars per initial client, after only the included costs. Same scale across all three sections. Solid blue: CRL; dashed grey: traditional.
| Month | Traditional monthly $ | CRL monthly $ | Traditional cumulative $ | CRL cumulative $ | Cumulative difference $ |
|---|---|---|---|---|---|
| 1 | 187 | 70 | 187 | 70 | -116 |
| 2 | 154 | 64 | 341 | 134 | -207 |
| 3 | 7 | 116 | 348 | 250 | -98 |
| 4 | 154 | 156 | 501 | 406 | -95 |
| 5 | 113 | 144 | 615 | 551 | -64 |
| 6 | 54 | 144 | 669 | 694 | 26 |
| 7 | 60 | 114 | 729 | 808 | 79 |
| 8 | 204 | 146 | 933 | 954 | 22 |
| 9 | 180 | 90 | 1,113 | 1,044 | -69 |
| 10 | 103 | 59 | 1,216 | 1,103 | -113 |
| 11 | 56 | 91 | 1,272 | 1,194 | -78 |
| 12 | 129 | 56 | 1,400 | 1,250 | -150 |
| 13 | 53 | 98 | 1,453 | 1,348 | -105 |
| 14 | 118 | 137 | 1,571 | 1,485 | -87 |
| 15 | 66 | 215 | 1,637 | 1,700 | 63 |
| 16 | 161 | 183 | 1,798 | 1,883 | 85 |
| 17 | 82 | 146 | 1,880 | 2,029 | 149 |
| 18 | 68 | 22 | 1,948 | 2,051 | 102 |
| 19 | 129 | 121 | 2,077 | 2,172 | 95 |
| 20 | 72 | 43 | 2,149 | 2,215 | 66 |
| 21 | 95 | 95 | 2,244 | 2,310 | 66 |
| 22 | 109 | 55 | 2,353 | 2,365 | 12 |
| 23 | 63 | 63 | 2,416 | 2,428 | 12 |
| 24 | 127 | 141 | 2,543 | 2,569 | 26 |
In the sample, CRL reduces the worst daily loss and maximum decline from peak, but increases the initial cash deficit. Counterparty risk on the internalised share and execution and liquidity risks on the external share remain.
| Measure based on realised cash flows | Traditional $ | CRL $ |
|---|---|---|
| Worst loss over a complete day | 132,981 | 115,233 |
| Maximum cumulative decline from peak | 315,338 | 199,737 |
| Maximum initial cash deficit | 51,148 | 148,381 |
Extremes across the 18 variants, not quantiles or capital requirements. Cash flows recognised at closes and modelled due dates: open exposures are not continuously revalued. Maxima may occur on different dates and in different variants and must not be added together. Definitions, trigger concentration and stress →
The actual internalised share, provider cost, capital, collateral and settlement timing. The cost calculator uses 10%; the economic calculator allows other shares.
Margins, cost components and a calculator for the broker's conditions.
The original configuration leaves positive modelled margin on average in all three books. This margin must cover acquisition, additional support, the actual provider price and capital: it is not yet profit after all costs. Aggregate risk measures do not confirm uniformly lower broker risk.
5× leverage, unchanged premium and trigger. No assumption of greater skill or loyalty among premium clients. Three book models applied to the same product; A-book execution of 5 seconds + 0.5 pip and provider at theoretical premium Q.
| Book | Average margin / client, 24 months | Minimum / maximum across variants | Positive margin |
|---|---|---|---|
| A-book | 2,127 $ | 1,725 / 2,449 $ | 18/18 |
| Hybrid 10% B-book | 2,569 $ | 1,845 / 3,291 $ | 18/18 |
| Pure B-book | 6,542 $ | -283 / 12,467 $ | 15/18 |
After commissions/operating costs, licence, funding and fixed costs under the book conventions; before the unquoted costs below. Correlated variants, not confidence intervals. Absolute margin is not the additional coverage cost compatible with parity against traditional leverage.
Arithmetic example, not a quote: average cumulative Q is $95,688 per client over 24 months. A 1% provider loading uses $957 of the $2,127 A-book margin, leaving $1,170 before other missing costs. This is below the traditional $1,555: parity is lost at 1%. Average parity is about 0.60% above Q (0.35–0.89% across variants), not zero CRL margin.
Zeros show margin without additional charges: they do not establish zero costs or full profit. Support beyond costs already included. Capital is allocated constantly and remunerated for 24 months; it is not calculated by the model. Provider loading applies only to the external share. No change to client fees.
Incremental contribution = CRL margin minus traditional margin forgone on the migrated share, minus additional costs. CAC applies only to new clients and transition costs only to migrated clients. A homogeneous client mix is assumed: this is not a commercial forecast. Taxes and other costs not entered remain excluded.
In the CRL B-book, margin is positive in 15/18 variants. The worst loss over a complete day is $621,961 versus $1,411,410 for traditional. Maximum cumulative decline from peak, however, is $5,513,276 versus $4,333,381. These are maxima along simulated historical portfolio paths of 1,000 clients, not quantiles or regulatory capital.
Different units: the economics table above is per client; the measures below are for a portfolio of 1,000 clients. Client deposits are never treated as the broker's own capital.
Initial deficit: maximum cumulative negative balance of modelled cash flows, with zero initial broker capital and retained profits. Decline from peak: maximum drawdown of the same balance. Cash flows are recognised at closes, execution adjustments and monthly fixed-cost dates; this is neither a mark-to-market account nor an actual treasury plan. Incomplete UTC days at the boundaries are excluded from the daily metric. Maxima across 18 variants; maxima may occur on different dates and in different variants and cannot be added together.
Only the positive incremental 4× payout on the externally hedged share, advanced until the provider's assumed reimbursement. Elapsed clock hours; reimbursement guaranteed at the end of the delay. Excludes return of deposits, margin calls, the underlying leg, prefunding and default. Must not automatically be added to the initial deficit. Counts are activations, not unique people.
Deterministic additional loss if all closes in the historical 1h window suffer the same adverse move on the 5× exposure. This is not a new price path, has no assigned probability and generates no new triggers. External share only: no hedge execution loss is charged to the pure B-book.
36 trade-by-trade ledgers, 18 pairs, 4 execution stresses and 3 book models. 10,368 monthly reconciliations to previously published results. Clients are aggregated at the same timestamps, preserving observed historical dependence. No independent sampling that removes concentration.
Still to be measured: economic and regulatory capital, mark-to-market open exposures, provider collateral and contractual liquidity, executable pricing and commercial costs by segment. Comprehension, willingness to pay, repeat use and retention require evidence from real clients. The model does not infer these from the “premium” label.
Internal checks on the extension; independent review of the baseline does not automatically extend to this module. Cost sensitivities do not change risk charts, which are calculated under the original conditions.
Under the disclosed economic conditions, CRL reduces loss severity relative to traditional 5× leverage. It does not reduce the number of loss-making clients in every cohort. For the broker, the outcome changes with how counterparty risk is managed.
24-month comparison: average of 18 correlated variants, original pricing, provider at theoretical premium Q, A-book stress of 5 seconds + 0.5 pip. Illustrative hybrid: 10% internalised, identical for both products. No retention benefit added to margins.
The commercial question: how much capital does CRL preserve, at what cost to the client and with what sustainable margin for the specific broker? The calculator also shows the conditions under which CRL produces less margin.
The A-book earns service revenue and transfers risk under the model's assumptions. In a pure B-book, the broker is the counterparty to client outcomes. The hybrid combines both components, with the same internalised share for both products: no client selection.
Economic sensitivity at fixed dates and notionals. Changing pricing here does not resimulate the client behaviour or wealth shown above. No cost of capital or retention benefit is automatically included.
Thresholds for one parameter at a time, all others unchanged; ranges across scenarios, not confidence intervals. Thresholds cannot be added together.
Margin after included costs only, before taxes and undisclosed costs. The pure B-book includes no external hedge, netting or risk capital. Risk remains with the broker even when the historical counterparty result is favourable.
A-book = commissions + service + additional spread + retained funding − operating costs − licence − fixed costs − execution deviation − provider loading − other costs.
Pure B-book, at disclosed pricing = initial capital − client wealth − operating costs − licence − fixed costs. The counterparty result contains premium, funding and payoff: the table breakdown avoids double counting. Pricing changes are applied to fixed trades.
Hybrid = (1 − internalised share) × A-book + internalised share × pure B-book. The share is identical for both products.
Available: monthly and cumulative margins, revenue and cost components, cross-model comparisons, economic thresholds, capital erosion and activity. Not estimated: LTV, CAC, LTV/CAC, acquisition payback, capital consumption and actual retention; the necessary broker data are missing.
Baseline assumptions, evidence and checks.
To avoid constructing a CRL advantage by artificially penalising its comparator, the traditional benchmark imposes discipline constraints. The dossiers also describe more aggressive behaviours; these are not combined here into a scenario designed to produce ruin.
No overlapping positions that multiply total account exposure. The effect of simultaneous portfolios is not attributed to the result.
No increase in leverage after a loss and no doubling down to recover losses. Notional is five times remaining capital: it decreases when the account loses value.
Stops and time-based closes are applied by the rules, without discretionary cancellation. Extensions of losing positions are capped; a loss is not left open indefinitely.
What this approach implies. These constraints limit the accumulation of exposure and some channels of extreme loss. For the same adverse move, greater exposure would amplify the loss. However, the number of additional accounts that would be depleted has not been measured: trades, costs and favourable outcomes would also change. The disciplined traditional benchmark does not represent the entire population described in the dossiers and is not a guaranteed statistical bound.
The comparison includes fixed activity and more frequent re-entry after a positive day. This behavioural transfer is disclosed; it is not a full calibration of retail clients. Dossiers CB01, CB02, CB05 and CB06 .
| Item | Actual model rule | Implication for interpretation |
|---|---|---|
| Exposure | Traditional: notional of 5× remaining capital. CRL: 1× before the trigger, payoff amplified to 5× at the trigger. One position per client at a time. | Lower downside exposure is a property of the product comparison; its benefit alone does not establish the added value of the premium and amplification. |
| Stops and targets | STD: stop / target of 5% of capital, equivalent to approximately ±1% in price before costs. CRL: stop of 5% in price and trigger z=0.35. | Contractual thresholds differ. The comparison does not match stop distances on the underlying; stops are not calibrated to individual volatility. |
| Durations | Declared profiles: 4 / 8 / 16 / 24 / 48 hours. Initial durations applied by index: 1 / 4 / 8 / 16 / 24 hours. Losing positions extended up to 120 hours; close on Friday at 20:00 UTC or earlier at the specified limits. | The mapping between profile and initial duration is explicit but not empirically calibrated. Time-based closes limit prolonged losses. |
| Behaviour | Fixed activity or increased re-entry probability after a win. The coefficient originally relating to volume is applied to entry probability. | A transferred mechanism, not a full calibration. No recovery of losses through higher leverage; after a loss, notional decreases with capital. |
| Churn | Voluntary churn disabled; technical remaining-capital threshold of $500 in the engine. No new clients, deposits or withdrawals in the model. | The 1,000 clients still trading do not constitute an empirical survival curve. Preserved capital is not automatically translated into retention. |
| Costs and broker | Annual funding of 3% of notional; commission of $20 per million per side under the disclosed conventions; CRL service fee of 2.75 bp; provider at theoretical Q. A-book broker with included costs only. | Funding favours the client comparison when exposure decreases. The broker's funding/spread share, the provider's executable price and cost of capital can reverse the advantage. |
| CRL execution | The client's net result is locked in at the trigger. Subsequent execution deviations fall on the broker. Four stresses applied to the same client paths. | Execution risk is transferred to the broker; the four stresses are not four independent client samples. |
The dossiers document erosion, repeated activity and churn in populations with different instruments, leverage and definitions. They do not provide joint calibration for our 5× EUR/USD profiles. The current model imposes duration, size, sequencing and exit policy: these constraints contribute to the outcome. The gap versus the sources cannot be corrected by choosing a target churn rate or loss.
For an adverse move d and an invested fraction f of capital, the gross pre-trigger loss is approximately f·d for CRL and 5·f·d for STD, plus their respective costs and the CRL premium. This is a mechanical relationship: it does not identify the distribution of d, chosen stops, trade frequency or survival.
The sources support the mechanisms of erosion, leverage, activity and churn. Populations, measures and periods are identified to avoid inappropriate numerical transfers.
In retail FX, the relative probability of stopping is associated with recent results; after a positive day, the hazard is 16% lower than after a negative day.
Use in the comparison: Supports the link between results and churn. Does not provide an absolute monthly rate to apply to CRL.
Primary source · pp. 11 and 17The US leverage restriction is associated with an estimated 23% reduction in volume and an approximately 40% reduction in losses among highly leveraged traders.
Use in the comparison: Supports the relevance of leverage and activity. Does not identify a probability of ruin at 5×.
Primary source · NBER 24176, June 2019 revision, p. 2In the Taiwan equity day-trader sample, survival is 44% at one year and 24% at two years, under specific entry and cessation criteria.
Use in the comparison: Documents persistence and exit. Does not measure capital depletion or a 5× leveraged Forex cohort.
Primary source · October 2017 manuscript, p. 12Average quarterly occurrences of negative balances fall from 21,727 to 2,524 following a package of CFD rules.
Use in the comparison: Documents extreme losses. These are aggregate occurrences, not ruin probabilities for unique clients; the intervention includes multiple measures.
Primary source · p. 11, para. 31Describes high retail-client turnover and some issuers' reliance on acquiring new clients.
Use in the comparison: Supports the analysis of lifetime and acquisition. Does not publish a monthly curve transferable to this simulation.
Primary source · p. 17, para. 61Highlights conflicts linked to frequency, volume, spreads and charges, and reports churn practices among a subset of firms.
Use in the comparison: Supports separating fees from the counterparty result. Does not attribute such conduct to all brokers.
Primary source · pp. 2–4Dossiers containing full texts and hashes are retained in the research file. Individual-level microdata needed to jointly estimate behaviour and churn for our profiles are unavailable.
Engine reconstructed through an independent implementation. 41,472 monthly KPI comparisons against trade ledgers; maximum residual below $0.000001. Hashes of 113 source files verified.
Accounting, path and cash-flow consistency verified. Three historical windows, one seed and correlated variants: the scale of the calculations is not out-of-sample validation or empirical calibration of retention.
The model quantifies a benefit in loss severity under the disclosed conditions. The broker's economic case depends on its book model and pricing. Both favourable and unfavourable results are part of the assessment.
Standalone English document: overview, A-book / B-book / hybrid, comparison discipline, charts, monthly KPIs, rules, checks and bibliography. The PDF uses the disclosed conditions and the average of the 18 variants; it does not incorporate temporary calculator changes.
This page presents quantitative research on the CRL model for brokers and other professional counterparties. It describes a product design and its simulated economics; it is not addressed to retail investors.
All figures are model simulations on historical EUR/USD data (2017–2022 windows) under the assumptions stated on this page. They are not observed client outcomes, forecasts or guarantees of future results.
Nothing here constitutes investment advice, a personal recommendation, or an offer or solicitation to buy or sell any financial instrument or service.
Leveraged trading carries a high risk of rapid capital loss. In these simulations most clients end the 24 months below their initial capital, under both traditional 5× leverage and CRL.
This page sets no cookies and loads no third-party scripts, fonts or analytics. It is hosted on GitHub Pages; GitHub may process technical data such as IP addresses to deliver and protect the service.
Published by: CRL Technologies, Inc. · 1313 N Market Street, Suite 5100, Wilmington, DE 19801, United States · info@crl-technologies.com