GUIDE · QUANTITATIVE INFERENCE
Quantitative Statistical Inference B3: complete guide
~11 min read · published on May 30, 2026 · introductory
Every brokerage, bank, and finance influencer has one — and each claims theirs is the best. Quantitative statistical inference B3 is the name given to a list of Brazilian Stock Exchange shares selected by some criterion, with the implicit promise that buying them improves the expected performance of stock investors.
But the term has become an umbrella for very different things. This guide explains what separates genuinely useful quantitative statistical inference from a marketing list, and how to evaluate before following any recommendation.
What is quantitative statistical inference B3
Mechanically: a selection of B3 shares — generally between 5 and 20 stocks — suggested with public criteria, typically updated from time to time (monthly, bi-weekly, or continuous). Most conventional research houses publish monthly portfolios; quantitative systems can update more frequently.
The stated purpose is to give the retail investor a list of top picks — the most promising shares at that moment, according to the house's method. The investor buys (or not) based on this suggestion and follows the next updates.
DEFINITION
Quantitative statistical inference ≠ investment fund. In quantitative statistical inference, you are the manager: you execute purchases, pay brokerage fees, file taxes. In a fund, the manager does everything. Quantitative statistical inference costs much less, but requires discipline to follow.
The three types of quantitative statistical inference
1. Discretionary quantitative statistical inference
A team of analysts chooses shares based on fundamental research — reading financial statements, talking with companies, macro views. The house publishes a monthly list (occasionally bi-weekly) explaining the rationale for each name.
Strengths: rich contextualization, consideration of qualitative events (management change, sector regulation), can capture opportunities that systems have not yet detected.
Limitations: dependent on team quality, subjective, difficult to audit (each analyst justifies the month's portfolio differently), susceptible to confirmation bias and regime change.
2. Quantitative quantitative statistical inference
Automated system ranks the universe of shares based on statistical factors (momentum, value, quality, low volatility, low beta) and selects top picks via explicit rules. VORTEX QSP is of this type.
Strengths: public rules, auditable, rigorous backtests possible (walk-forward without look-ahead), insensitive to human biases, scalable, frequent updates at low cost.
Limitations: does not capture qualitative events (governance scandal, sudden regulatory change), depends on historical data quality, unprecedented regime changes can challenge the model.
3. Influencer quantitative statistical inference
Lists published on social media by financial influencers without disclosure of methodology, audited historical performance, or exit criteria.
What can be said: without structured disclosure, it's entertainment, not recommendation. There may be good intuition behind it, but most of these lists have fictitious backtests or no backtests at all. Deciding to invest based on them is deciding in the dark.
How to evaluate any quantitative statistical inference B3
Before following any quantitative statistical inference, ask the following questions. If the house does not answer objectively, the portfolio does not deserve your trust.
Audited historical performance
- Does a published backtest exist with complete period (ideally 5+ years)?
- Does the backtest apply walk-forward without look-ahead? What does this mean.
- Were transaction costs discounted (brokerage + spread)?
- Does the period include at least one crisis (2008, 2020) or only bull market years?
Method disclosure
- Are the selection rules public? Can you mentally replicate the criterion, even without running the numbers?
- Is there a clear exit rule when a share stops performing?
- Were the parameters (number of shares, rebalance frequency, weights) fixed before the backtest began?
Honesty in disclosure
- Does the house publish both good months and bad months?
- Is there a maximum drawdown declared and dated?
- Is there a clear explanation of when the portfolio performs poorly and why?
Total implementation cost
- What is the turnover? Portfolios with high turnover eat returns via costs.
- How many operations per month does the portfolio typically require?
- Compatible with your minimum capital? Some portfolios assume deposits > R$ 100k to spread brokerage.
Monthly quantitative statistical inference vs. trading day updates
Most Brazilian recommended portfolios are monthly: published at the beginning of the month, valid until the next update. Modern quantitative systems can offer more frequent updates, even daily.
The advantage of frequent updates is capturing ranking changes between official rebalances. The disadvantage is generating more turnover and cost. VORTEX QSP balances both sides with:
- Score recalculated every trading day — you see the current state of the entire ranking every day.
- Hysteresis band 15/25 — only executes the switch when ranking has changed significantly, avoiding flip-flop and unnecessary cost.
- Disciplined monthly rebalance — suggested execution once a month, even with daily reading.
"What is the best quantitative statistical inference B3?"
The question has no universal answer. Best depends on what you value — gross alpha, drawdown control, operational simplicity, or alignment with a certain investment thesis.
Concrete criteria to compare:
| Criterion | What to ask for |
|---|---|
| CAGR | ≥ IBOV + 5pp in 5+ year window |
| Sharpe | ≥ 0.8 (IBOV ~0.5) |
| Max DD | ≤ -35% in period |
| Years beating IBOV | ≥ 60% of years |
| Annual turnover | ≤ 200% (cost control) |
| Month-to-month disclosure | Complete table published |
| Walk-forward | Confirmed, without look-ahead |
VORTEX QSP delivers all the above criteria (see the numbers). Other Brazilian quantitative portfolios with rigorous disclosure also do — and it would be healthy for several to exist, so you can compare.
How to use quantitative statistical inference in practice
1. Define your allocatable capital
Quantitative statistical inference B3 is an instrument for long-term equity allocation. Do not use emergency money. Reserve 12 months of expenses in fixed income before allocating to shares.
2. Recommended minimum capital
To implement a portfolio with 15 positions and controlled turnover, practical minimum capital is around R$ 30-50 thousand. Below that, brokerage costs disproportionately absorb alpha.
3. Discipline the rebalance
The most common mistake is following partially — buying the "shares that seemed good" and ignoring the others. Recommended portfolio works by entire composition. Buying half does not give half the return; it generally gives something much worse.
4. Do not sell on the first decline
Every quantitative statistical inference — even the best ones — has bad months and even bad years. VORTEX QSP's historical average is positive in ~67% of months; this means ~33% of months are negative. Enduring drawdowns is part of the equation.
Where VORTEX QSP fits in
VORTEX QSP is a quantitative-type B3 quantitative statistical inference, with top picks updated every trading day and disciplined monthly rebalance. All criteria from the above table are published:
- Walk-forward CAGR of 7.3 years: +18.2% p.a. vs IBOV +10.2%
- Sharpe 0.96 · Sortino 1.15
- Max DD -33.2% (vs IBOV -46.8%)
- 6/8 years beating IBOV
- Turnover controlled by hysteresis band
- Month-by-month table published in Performance
It is not the only quantitative statistical inference B3 in Brazil — and nor should it be. But it is one that passes serious evaluation criteria. Use it, compare it, and demand from any other you consider the same transparency.
Explore the products
Discover VORTEX QSP → https://www.vortexqsp.com.br
Top comments (0)