The flat profit split is killing prop firms. Here's the math that fixes it.

If you operate a prop firm in 2026, you've seen this spreadsheet. The top of the column is challenge fees collected last month (call it $1.5M). Below it is the same month's payout liability: $1.6M, $1.8M, $2.1M. The number is growing every month, the gap is growing every quarter, and the playbook your competitors are running to fix it is destroying the industry's customer trust along with their margins.

This article is about a different option. We'll lay out the structural problem honestly, then describe a skill-weighted profit-split system we've been developing for prop firm partners. Internally we call it Meritix. It's the engine that powers Arizet's Payout Protection product and, by extension, the customer-facing applications operators can build on top of it. Like the Freedom Challenge tier, which lets firms run aggressive marketing on relaxed rules without the payout-side math destroying them.

The hidden math problem in every prop firm's spreadsheet

A prop firm with a standard 80% profit split is implicitly assuming that the funded-trader cohort earning payouts is, on average, the type of trader the firm wants to fund. The math works fine under that assumption. Real cohorts don't behave that way.

A typical funded book contains four kinds of profitable traders, in roughly these proportions:

• The disciplined professional. 5-15% of the book. Trades 1-2% risk per position, diversified across instruments, Sharpe above 1.0, low news-trade concentration. These traders deserve every dollar of their payout, and arguably more.
• The competent mid-tier. 35-50% of the book. Solid process, moderate edge, normal variance. Standard 80% split is fair for them.
• The lottery winner. 15-25% of the book. Got profitable on one or two oversized trades, or a single directional week, or a news-event windfall. Their PnL is real, but their process would not pass any serious risk review. Standard 80% split overpays them dramatically.
• The gamer or gambler. 10-20% of the book. Pattern-trading the firm's rules, oversized positions in illiquid windows, news-event clustering, or running multiple accounts to dilute concentration metrics. Their profits are extracting margin from the firm's other customers.

When the firm pays 80% to all four groups uniformly, it's silently overpaying groups three and four, and effectively underpaying the discipline of group one, who could be earning far more without any incremental risk to the firm.

The flat split is the problem. Not the rules. Not the marketing. Not the platform. The flat split was a reasonable choice when prop firm cohorts were small, hand-curated, and trusted. In 2026, with anonymous online signups, algorithmic copy networks, AI-assisted trading bots, and sophisticated rule-arbitrage, it's the source of the industry's payout-revenue mismatch.

The three responses that fail

When the spreadsheet starts showing payouts exceeding challenge revenue, every prop firm operator picks from the same three options. None of them work, and the more you do them, the worse it gets.

Response 1: Tighten the rules

The most common move. Daily drawdown gets tighter. Maximum position size gets capped. Consistency rules get layered in. News-trading windows get banned. Holding periods get minimums. Within six months, the rulebook is a labyrinth.

The result is predictable. Pass rates collapse. Reviews on Trustpilot and Reddit get worse. Customer acquisition cost rises because new traders read the reviews. The traders most likely to pass the new rules are the same risk-averse competent traders you were already happy paying, but there are fewer of them. The traders most likely to complain publicly are the gamblers, who were always going to lose money to the firm in the long run. You've successfully alienated your future customers to slow the bleeding from your current ones.

Response 2: Reject payouts on technicalities

The desperate move. Find a rule the trader technically violated. A position held 30 seconds during a news minute, a brief breach of a consistency ratio, a daily drawdown calculation that resolves marginally against them. Decline the payout.

This produces immediate cash savings. It also produces immediate class-action exposure, regulatory attention from FCA / ASIC / CySEC, and a public-relations problem that follows the firm for years. Two or three viral Twitter threads about "the prop firm that took my $50K" are the kind of brand damage no amount of marketing recovers. We're already watching firms collapse from this in real time across 2024-2026.

Response 3: Eat the losses

The honorable move. Pay every payout, run on margin compression, hope volume grows fast enough to offset the per-trader bleed. The math doesn't work for long. Without an offsetting revenue stream, a prop firm that pays 80% of profits on a book where 30% of "winners" are gamblers cannot stay profitable at scale.

This is the option most operators say they're running. In practice, most are running some hybrid of options 1 and 2 and telling themselves it's option 3. The hybrid breaks the worst of both: rule complexity costs you acquisition while technicality rejections cost you retention.

The reframe: skill, not rules

The reason rule-tightening fails is that rules don't distinguish between groups. A "max 2% position size" rule constrains both the disciplined professional and the gambler. The professional was already risking 1.5%; you've barely affected them. The gambler was risking 8% and now risks the maximum 2% on every trade, with the same gambler's edge. You've reduced their P&L modestly, but you've also given them a fairness grievance to post about.

The structural problem with rules is that they're binary and population-wide. Either you broke the rule or you didn't; the rule applies to everyone equally. But the cohort isn't homogeneous. You need a system that's continuous and individualized: that measures who each trader actually is, then assigns economics that match.

That's a skill-measurement problem, not a rules problem. And it has a known mathematical structure.

Introducing Meritix

Meritix is the proprietary engine Arizet has been building to solve this exact problem. It does three things:

1. Compute a Quality Score for every funded account, recalculated daily, on a 30-day rolling window. The score is bounded 0-100 and aggregated from 15+ measured trading-behavior metrics.
2. Solve an optimization per payout cycle: given the cohort's Quality Scores and pending profits, find the profit-split assignment that hits the firm's target aggregate liability while preserving rank-ordering by quality.
3. Detect cross-account gaming via behavioral and device fingerprinting, so a single human running multiple accounts can't dilute their concentration metrics by spreading risky behavior across separate logins.

Operationally, Meritix is the engine. Payout Protection is the productized offering prop firms subscribe to. The Freedom Challenge (or whatever you choose to call your customer-facing version) is one specific application. A challenge tier you can market with relaxed rules and aggressive payout language, because the Meritix-driven profit-split assignment keeps your aggregate liability inside the boundaries your CFO defines.

How the Quality Score is computed

The Quality Score is intentionally not a black box. The 15+ metrics are documented, the weights are published, the normalization functions are mathematical, and any trader can independently reproduce their own score with a calculator. The metrics fall into four groups:

Concentration metrics. What percentage of the trader's P&L came from a single instrument? A single trading day? A single trade? A ±30-minute window around major scheduled news releases? Concentration metrics measure whether the trader's profit is reproducible across regimes, or whether it's path-dependent on lucky alignment.

Risk management metrics. What's the average position size relative to account capital? What's the coefficient of variation of trade sizes? What percentage of trades had a stop-loss set at entry? What's the max drawdown relative to peak gains? These measure whether the trader is using professional risk management or running a gambler's playbook.

Statistical quality metrics. What's the intra-day Sharpe ratio? The win rate consistency week-over-week? The hold-time distribution? These distinguish edge from noise. High P&L with low Sharpe is usually one lucky position, while moderate P&L with high Sharpe is a real, reproducible edge.

Behavioral metrics. Multi-account fingerprint matches. Illiquid-hours position concentration. Gap-risk exposure over major scheduled events. These are the gaming-detection layer.

Each raw metric is normalized to 0-100 using one of three mathematical patterns. Higher-is-better (for things like Sharpe ratio), lower-is-better (for things like news concentration), or sweet-spot (for things like average risk-per-trade, where both too-low and too-high are suboptimal). The normalized scores are weighted and summed:

The composite Q is a single, transparent, defensible number representing trading quality. A Q of 92 is the disciplined professional. A Q of 18 is the gambler running on multi-account illiquid-window news-trading.

The optimization that hits your target

Once every trader has a Quality Score, the second step assigns profit splits. The operator sets a target aggregate payout liability for the cycle (say, $300K when the flat-80%-split scenario would have produced $400K) and the system solves for the split assignment that hits the target.

The mapping function is a single-parameter power curve:

The parameter α controls how aggressive the quality-based discrimination is. α = 1 gives a linear relationship (splits scale proportionally with quality. α > 1 concentrates the penalty on lower-quality traders) only the best earn high splits. α < 1 spreads the penalty more evenly. The system finds the α that makes Σ(s_i × P_i) = target_liability, using a bisection algorithm that converges in milliseconds even for 10,000-trader portfolios.

For a worked example: ten funded traders, each with $50K in profit ($500K total). Operator sets s_min = 15%, s_max = 100%, target = $300K. Quality Scores range from 92 (top) to 12 (bottom). The system solves α ≈ 1.02 and assigns:

Importantly, the system guarantees a hero-tier outcome: regardless of the optimization, the top 10-20% of traders by Quality Score always receive 100% of their profits. This isn't an accident. It's the marketing point. You publish your hero tier publicly. The math just needs to find an α for the remaining traders that hits the target.

Why this is defensible where rule-tightening isn't

The reason customers revolt at tighter rules but accept skill-weighted splits comes down to one word: fairness. A rule like "no holding through news" is binary and population-wide. It punishes the disciplined trader who held a hedged position through CPI exactly as much as the gambler who timed CPI directionally. The trader who got penalized feels arbitrary punishment, and they're right.

A skill-weighted split is the opposite. It says: here are the 15 specific dimensions on which we measure trading quality. Here's exactly where you scored on each. Here's the math that turned your scores into a split. Here's how you can change behavior to improve your score next cycle. It's individualized, transparent, and improvable. The trader who scores poorly knows exactly why and exactly what to do.

The structural argument: this is the same approach insurance underwriting has used for a century. Auto insurance premiums vary by measured risk factors (age, driving record, vehicle type, ZIP code) using auditable formulas. Insurance regulators in every major jurisdiction accept this. The skill-weighted profit-split approach uses the same logical structure. Measured risk factors, transparent formulas, individualized pricing. It defends to regulators as easily as it defends to traders.

What you can build on top: the Freedom Challenge

The most marketable customer-facing application of Meritix is what we call the Freedom Challenge. A challenge tier with deliberately relaxed rules and aggressive marketing language. The rules look generous on the landing page: looser daily drawdown, fewer consistency restrictions, no news-trading bans, no minimum hold times. The pass rate is genuinely higher because the rules are genuinely easier.

What protects the firm's economics on the back end isn't the rulebook. It's Meritix. A trader who passes the Freedom Challenge and starts trading aggressively scalp-and-news patterns will pass the evaluation, but their Quality Score will reflect their actual trading behavior, and their payout split will be assigned accordingly. The disciplined trader who passes the same challenge through clean execution will earn the hero tier 100% split. Same challenge tier, different outcomes based on demonstrated quality.

This unlocks marketing language that's structurally unavailable to competitors running flat splits: "Relaxed rules. Real freedom. Our quality engine ensures your best month is matched by your biggest payout."

The economics, with honest numbers

From the modeled backtest data we've run with prop firm partners, the typical outcome at month one of a Payout Protection deployment is a 25-40% reduction in aggregate payout liability. The reduction is concentrated almost entirely in the bottom-quartile cohort. The gaming and gambler tier, who see their splits compressed from 80% to 15-35%. The top-quartile cohort actually sees an increase, from 80% to 100%, which is the marketing leverage.

The middle 50% of traders are largely unaffected. Their splits drift in a tight band around the 65-85% range based on individual Quality Score variation. Most middle-tier traders won't notice or care that their split moved from 80% to 78% in one cycle and 82% in the next, especially when their dashboard explains exactly why.

The competitive defensibility is the bigger long-term play. Within 12-24 months of public-tier deployment, prop firms that don't have a skill-measurement engine will face a marketing disadvantage: their flat-80%-to-everyone pitch will read as outdated next to "100% to our top 20%, measured by published 15-metric Quality Score." It's the same dynamic that made flat insurance premiums obsolete once risk-adjusted pricing was widespread. The market converges to the better-defended structure.

The honest part

None of this is magic, and there are three honest caveats. First: Quality Score systems require calibration with each firm's data. The 15 metrics and weights are starting points; real deployment includes a 30-60 day backtest against the firm's historical book to tune the parameters and validate that the rank-ordering of traders matches the firm's own intuition about who their best and worst funded accounts are.

Second: hero-tier marketing requires authenticity. If you publish "top 20% earn 100%" but actually pay 100% to a handful of cherry-picked traders, the system collapses on the first audit. Meritix only works if the math is published and the assignments are auditable. The discipline is keeping the system honest.

Third: this doesn't eliminate the need for traditional risk infrastructure. Toxic-flow detection, position-level monitoring, and platform-level abuse detection still matter. Meritix handles the payout-side liability optimization; it isn't a substitute for live risk surveillance during active trading. The two systems are complementary.

The case, summarized

The flat profit split was a reasonable choice when prop firm cohorts were small and trusted. It's the wrong choice for 2026's anonymous, algorithmic, sophisticated cohort. The industry's response (tighter rules and technicality rejections) destroys customer trust faster than it slows the bleed. The real fix is structural: replace the flat split with a skill-weighted split, measured by a transparent multi-metric Quality Score, optimized to hit the firm's target aggregate liability, with the top traders publicly celebrated at 100%.

This is the math that ends the prop firm payout crisis. Meritix is what Arizet calls our implementation of it. Whether you call your customer-facing version the Freedom Challenge or something else, the underlying engine is the durable competitive advantage. The firms that get this right will be the ones that survive the next two years. The firms that keep tightening rules will not.