Money Management is the last link in the Portfolio Rules chain: it’s what turns a list of selected strategies into a portfolio with real weight. After filters, the Correlation stage and the Rebalancer have decided which strategies to keep, Money Management decides how much of each: the size to give it — the number of contracts, or the corresponding monetary amount — and therefore how much risk to put on it.
In simple terms: selecting the right strategies is half the job; the other half is sizing them so that none, alone or together, can cause disproportionate damage. QANTHOS’s Money Management module is built entirely around this idea of risk control: it starts from a risk budget per trade, translates it into contracts, and then lines up a series of safeguards — position limits, correlation adjustments, joint-loss management — that keep the portfolio from concentrating risk where you don’t want it.
An important premise. QANTHOS is an analysis, validation and discovery tool, not a financial advisor. The methods, risk percentages and parameters described in this chapter are configurable levers and educational starting points, not recommendations on how much to risk. The thresholds cited (the “2% rule”, a Kelly fraction, a Safety Margin) are reference conventions common in systematic trading literature: useful as orientation, not as prescriptions. How much risk is right for you depends on your capital, costs, broker and tolerance: you are always the final judge.
1. Where it lives and how it’s structured
Money Management lives in two complementary places. In the Portfolio Rules header there’s a compact panel with a preset selector and a synthetic display (capital, method, risk percentage, any active adjustments). From there, the ⚙️ Money Management button opens the full configuration dialog, organized into five tabs:
| Tab | What it’s for |
|---|---|
| Method | Choose the sizing method (how much to risk is calculated) and its parameters. |
| Base | Initial capital, worst-loss estimation method, safety margin. |
| Correlation Penalty | Adjust sizes based on how much strategies move together. |
| Trade Overlap | Adjust sizes based on how much strategies operate on the same days. |
| Limits | Caps and floors: max per strategy, min and max contracts. |
The compact panel and the dialog are the same configuration seen at two levels of detail. The presets (Conservative, Moderate, Aggressive, Half-Kelly, Custom Balanced, Custom) let you start from a coherent base with one click; as soon as you touch a parameter in the dialog, the preset automatically switches to “Custom” — a sign the configuration is now yours.
2. The heart of sizing: from risk to contracts
Before the methods and tabs, it’s worth understanding the base mechanism, because this is where risk control originates. The most transparent method — Fixed Fractional — sums it up in a single formula:
contracts = (Capital × Risk%) / (Worst Trade × Safety Margin)Read it like this: you decide how much of your capital you’re willing to lose on a single trade (the Risk%), you estimate how much that strategy can lose in one shot (the Worst Trade), you add a prudential buffer (the Safety Margin), and the number of contracts follows. The bigger the estimated “worst hit”, the fewer contracts the strategy receives — automatically.
The important detail, from a risk standpoint, is that QANTHOS doesn’t just compute contracts: it also applies a final constraint guaranteeing that the potential loss of the chosen number of contracts — contracts × |Worst Trade| × Safety Margin — never exceeds the declared risk budget. It’s a double net: sizing targets the budget, and the constraint verifies it isn’t exceeded.
Why three ways to estimate the “Worst Trade”. The worst loss isn’t an obvious number: the single worst trade ever might be a one-off event, while a percentile captures a “typical bad trade”. That’s why the Base tab offers three choices, each with a different idea of prudence, described below. It’s one of the most consequential risk-management decisions: estimate the loss too low and you oversize; estimate it too high and you leave size on the table.
All of this is recalculated at every rotation: as the pool evolves, sizes adapt too, and with them the portfolio’s risk profile over time.
3. Sizing methods (Method tab)
The Method tab is where you choose the sizing philosophy. QANTHOS groups methods into three families. When you select a method, only its own parameters appear (for example Risk % for Fixed Fractional, the variant for Kelly), and a short description summarizes its behaviour.
| Method | Family | Underlying idea (risk-wise) |
|---|---|---|
| Fixed Fractional | Literature | Risks a fixed percentage of capital per trade. The transparent standard: per-trade risk is explicit and constant. |
| Kelly Criterion | Literature | Targets optimal geometric capital growth. The Half variant (default) halves full Kelly’s aggressiveness, a widely used prudential choice. |
| Fixed Ratio | Literature | Increases contracts only after accumulating a certain profit (Delta). Protects the initial capital and scales slowly with gains. |
| Percent Volatility | Literature | Normalizes risk by the strategy’s volatility: more volatile ones get fewer contracts. Equalizes the “bite” of each. |
| Optimal f | Literature | Maximizes geometric growth; very sensitive to the historical worst trade. Typically used with a fraction (e.g. 50%) to contain its aggressiveness. |
| CPR (Compound Position Risk) | Literature | Caps overall position risk, with optional profit compounding. |
| Equal Weight | Allocation | Equal weights (1/N). Simple and robust, an excellent starting benchmark. |
| HRP (Hierarchical Risk Parity) | Allocation | Allocates risk using hierarchical clustering of correlations. Robust, doesn’t require inverting the covariance matrix. |
| Risk Parity | Allocation | Equalizes each strategy’s risk contribution: more volatile ones weigh less. |
| Inverse Volatility | Allocation | Weight inversely proportional to volatility. A simple alternative to Risk Parity. |
| Min Variance | Allocation | Minimizes total portfolio variance. Can concentrate on a few low-volatility strategies. |
| MR Method (Cluster-Based) | Custom | Proprietary method: groups strategies into correlation clusters and balances risk across clusters, rewarding decorrelated ones and penalizing correlated ones. |
Two approaches, two ideas of risk. The literature methods reason strategy by strategy (how much to risk on this strategy). The allocation methods reason about the set (how to distribute risk across strategies), and for this they expose portfolio-level parameters such as Target Volatility and max/min weight per strategy. There’s no “right” choice: it depends on whether you want to control risk bottom-up (per trade) or top-down (per portfolio).
Allocation methods, when selected, show a dedicated parameter box: Target Volatility (the annualized volatility the portfolio aims for), Max Weight and Min Weight per Strategy (the weight limits that avoid both concentration on a few strategies and unusable crumbs).
4. The Base tab: per-trade risk parameters
The Base tab gathers the three ingredients that feed the risk formula.
| Parameter | Risk-wise meaning |
|---|---|
| Initial Capital | The reference capital on which risk percentages are computed. |
| Worst Trade Method | How to estimate the worst loss: 95th Percentile (a “typical bad trade”), 99th Percentile (more conservative — looks at the tail), Absolute Worst Trade (the single worst trade ever seen, the strictest). |
| Safety Margin | A multiplier (1.0–2.0×) applied to the worst loss. Higher = more buffer = fewer contracts. The general-prudence dial. |
Worst Trade Method and Safety Margin work as a pair: the first decides which loss to fear, the second how much margin to add on that estimate. Together they determine how conservative the entire sizing is, at a given risk percentage.
5. Portfolio-level risk management functions
This is where the module’s added value lies. Base sizing looks at each strategy on its own; but a portfolio’s real risk arises from how the strategies interact. QANTHOS offers three adjustment families — plus position limits — that step in after the base contract calculation to correct course. They’re the heart of what makes Money Management a risk-control tool, not just a size calculator.
A clarification that avoids confusion. In the main chapter you met the 🔗 Correlation stage, which excludes strategies that are too correlated. Money Management’s adjustments are different: they don’t exclude, they resize. The Correlation stage decides who’s in; Money Management decides with how much weight. They can be used together — first decorrelate the group, then dose the survivors’ sizes — and they answer two distinct questions.
5.1 Position limits (Limits tab)
The most direct safeguards, the boundaries every size must stay within.
| Limit | What it protects |
|---|---|
| Max % per Strategy | Cap on the share of capital a single strategy can absorb. Prevents one strategy from dominating the portfolio. |
| Min Contracts | Minimum contracts per strategy (default 1). Set to 0 it takes on a special risk-control meaning: it excludes strategies whose per-contract risk already exceeds the budget — better not to trade them than to trade them undersized. |
| Max Contracts | Absolute cap on contracts per strategy. A brake against out-of-scale sizes. |
5.2 Correlation Penalty — dosing by outcome (Correlation Penalty tab)
This tab adjusts sizes based on how much strategies win and lose together. It’s a unified tab with an enable switch and a selector choosing which of two systems to use — they’re mutually exclusive.
“Classic Correlation” system. Measures return correlation between strategies and acts in two directions: decorrelated strategies (correlation below the low threshold) get a contract bonus, because they add diversification; correlated strategies (above the high threshold) get a penalty, because they concentrate risk. Parameters:
| Parameter | Meaning |
|---|---|
| Correlation Frequency | Which P&L to compute correlation on: Daily (granular, noisier), Weekly (more balanced), Monthly (smoother). |
| Decorrelation Bonus | How much to reward decorrelated pairs (in %). |
| Correlation Penalty | How much to penalize correlated pairs (in %). |
| Low / High Threshold | The two thresholds defining “decorrelated” (below Low) and “correlated” (above High). |
| Period (months) | The window over which to compute correlation. 0 = all available history. |
“Co-Loss Risk” system. The subtler safeguard, and worth paying attention to. Two strategies can have low return correlation yet tend to lose on the same days: average correlation doesn’t capture this, but drawdown does. Co-Loss measures precisely the frequency with which two strategies lose together, and reduces contracts for dangerous pairs.
Its logic is “target-based”, and it’s worth understanding so it doesn’t surprise you:
| Parameter | Meaning |
|---|---|
| Target Risk % | The portfolio-level Co-Loss Risk threshold to stay under. The algorithm reduces high-Co-Loss pairs only while overall risk remains above this target. If the portfolio is already under it, nothing is reduced. |
| High Threshold % | Defines which pairs count as “high-risk” (those losing together more often than this threshold). Used to choose which to reduce first. |
| Floor Action | What to do when a strategy is already at the minimum (1 contract) but remains risky: Penalize Others (reduce the other side of the pair), Accept Risk (keep it with a warning), Exclude Strategy (remove it). |
| Period (months) | Calculation window. 0 = use the rotational filter’s period. |
Why changing Co-Loss sometimes “does nothing”. Since the reduction only kicks in when portfolio risk exceeds the Target Risk %, if you set a target higher than the already-observed Co-Loss, the adjustment never intervenes — and that’s the correct behaviour. To actually trigger it, the target must be set below your portfolio’s Co-Loss Risk. After generation, QANTHOS warns you if the adjustment was enabled but never bit, so you’re not left wondering why the results don’t change.
5.3 Trade Overlap — dosing by timing (Trade Overlap tab)
Trade Overlap is independent of the two previous systems and measures a different dimension: how much two strategies operate on the same days. It’s a matter of timing, not outcome.
The distinction matters and is worth keeping in mind: correlation is about whether strategies win or lose together (the outcome); Trade Overlap is about whether they’re in the market on the same days (the timing). Two strategies can have low correlation but high overlap — different results, but margin committed all on the same day. The adjustment rewards with a bonus strategies covering different days (more temporal diversification) and penalizes those concentrated on the same days. The only parameter is the Adjustment Weight (0.1–1.0), which doses the effect’s intensity.
When Trade Overlap is particularly useful. If you need to lower the portfolio’s required margin, this adjustment is a direct lever. Trades overlapping on the same days commit margin all at the same moment and push up the simultaneous margin peak; by penalizing high-overlap strategies and rewarding those covering different days, Trade Overlap spreads trades over time and tends to lower the required peak margin. A valuable consideration when available capital or broker margin requirements are a real operational constraint.
5.4 Three complementary lenses
Worth fixing the distinction, because it’s easy to confuse the three adjustments:
| Adjustment | What it measures | Question it answers |
|---|---|---|
| Correlation (Classic) | Return correlation | Do they win and lose together? |
| Co-Loss Risk | Joint-loss frequency | Do they lose together in the same periods? |
| Trade Overlap | Coincidence of operating days | Are they in the market at the same time? |
They aren’t strictly alternatives (Trade Overlap coexists with one of the two Correlation Penalty systems): they’re different lenses on the same problem — don’t put too many eggs that break at the same moment.
5.5 Room to maneuver: why starting size matters
There’s a practical condition that determines how much these systems can actually bite, and it’s worth keeping in mind: contracts are integers. An adjustment is expressed as a percentage — a bonus or penalty of, say, 20% — but the system can only add or remove whole contracts. This creates a huge difference depending on how large the sizes in play are.
If the portfolio is made of strategies with small sizes — 1, 2, at most 3 contracts — there’s almost no room to maneuver. A 20% penalty or bonus simply isn’t achievable: the smallest move the system can make is one contract. And adding one contract to a strategy that had only one isn’t worth a 20% reward, but 100% — the size doubles. The same applies in reverse: taking “a little” off a size of 1 means zeroing it out. In this regime the adjustments become coarse, or effectively inapplicable: the system has no room to dose.
When sizes in play are larger instead — 5, 6, 7 contracts or more — fine granularity becomes available again. Adding or removing one contract on a size of 6, for example, is roughly a 17% adjustment, close to the 20% required: the system can intervene surgically, genuinely distributing risk and giving the portfolio real balance instead of jumps.
It’s worth noting this limit only concerns instruments traded in whole contracts — typically futures, where size can only move in steps of one contract. For instruments allowing much more granular sizing — such as forex or CFDs — the problem doesn’t arise at all: size is finely adjustable, the systems can apply bonuses and penalties at practically the exact percentage required, and they intervene far more effectively, producing a decidedly more balanced portfolio.
The practical consequence. Risk-control adjustment systems — Correlation Penalty, Co-Loss, Trade Overlap — express their full value when sizes are large enough to be modulated. On portfolios with very small sizes the same systems, while active, have little room to act, and their effects can be coarse or nearly imperceptible. Worth keeping in mind when assessing how much these tools can matter in your case: the larger the sizes, the finer — and more effective — risk control becomes.
6. Portfolio-aware allocation as risk control
The allocation methods seen in the Method tab (HRP, Risk Parity, Inverse Volatility, Min Variance) deserve a separate note from a risk perspective: instead of starting from a per-trade budget, they start from how risk is distributed across strategies. HRP, in particular, uses hierarchical clustering of correlations to allocate without having to invert the covariance matrix — a robust approach when there are many strategies and correlations are unstable. They’re the right tool when you want the portfolio’s risk structure to drive the weights, rather than a uniform per-trade rule.
7. Common issues and how to avoid them
I set a high risk but sizes stay small. The Worst Trade Method or the Safety Margin are probably estimating very large worst losses: the formula’s denominator grows and contracts shrink. Check the consistency between risk percentage, loss estimation method and margin.
A strategy never enters the portfolio with sizing. With Min Contracts at 0, strategies whose per-contract risk exceeds the budget are deliberately excluded. If that’s not the behaviour you want, set Min Contracts back to 1 — accepting potentially under-budget sizes.
Changing Correlation Penalty or Co-Loss settings doesn’t change the results. It can happen that the adjustment is enabled but has no conditions to act on (no pair over threshold, or risk already under the Target). QANTHOS flags it with a warning at the end of generation, distinguishing “wasn’t needed” from “wasn’t computable”. For Co-Loss in particular, check that the Target Risk % is below the observed Co-Loss.
I enabled both Classic and Co-Loss. Not possible: the Correlation Penalty tab is unified and keeps only one active. Trade Overlap, instead, is independent and can coexist with either.
8. An example workflow
An exploration path, not a recipe. The choices depend on your capital, your instruments and your risk tolerance.
- Open ⚙️ Money Management from the Portfolio Rules header.
- In the Method tab, choose the sizing method and set its parameters (for Fixed Fractional, the Risk % per trade).
- In the Base tab, set the capital, choose the Worst Trade Method and the Safety Margin that reflect your prudence.
- In the Limits tab, define the boundaries: Max % per Strategy, Min and Max Contracts.
- If you want sizing to account for interactions between strategies, enable the adjustments in the Correlation Penalty and/or Trade Overlap tabs, keeping in mind Co-Loss’s “target-based” logic.
- Confirm, go back to Portfolio Rules and press GENERATE PORTFOLIO. Analyze the result in the Equity tab and, for Money Management detail, in the MM Analysis Tab.
- Change a parameter, regenerate, compare: this is how you understand how each risk lever affects the overall equity.
Money Management is the part of the system that translates good intentions about risk into concrete numbers. But it remains a tool: it sizes according to the rules you give it, not according to what’s “right” for you. Validating a sizing configuration — with Walk-Forward Analysis, stress testing and observation of real drawdown — is a step the tool facilitates but doesn’t replace.
9. To learn more
- Portfolio Rules — the rule chain and where Money Management fits in.
- Correlations — the final decorrelation stage, which excludes (distinct from Money Management’s size adjustments).
- Portfolio Rebalancer — concentration balancing, upstream of sizing.
- Tools Tab — per-ticker operating thresholds, underlying cost estimation.