How It Works

• A momentum approach looks for assets whose recent price movement suggests continued directional pressure, rather than trying to predict a reversal. It typically combines a measure of recent price change with confirming signals -- such as trading volume -- before treating that momentum as significant enough to act on.
• Entries are conditioned on multiple signals agreeing simultaneously, not any single measurement crossing a threshold in isolation.

When It May Work

• Momentum approaches have historically tended to perform better in trending market regimes, where directional moves persist over multiple periods rather than immediately mean-reverting.
• They tend to underperform in choppy, range-bound conditions, where short-lived moves reverse before a momentum-based entry can benefit from them.

Main Risks

• Momentum can reverse sharply and without clear warning, particularly around news events or thin-liquidity conditions -- a strategy that only looks backward at price has no way to anticipate that in advance.
• By construction, momentum entries occur after a move has already started, meaning some of the move is typically missed and the strategy is exposed to a reversal risk the earliest participants in the move are not.

What the System Analyzes

• Recent price change over multiple lookback windows
• Trading volume relative to its recent average
• Data freshness and confidence, so stale or unreliable data is never treated as a valid signal

How It Works

• Trend-following systems use indicators built from price history -- commonly moving averages of different lengths -- to characterize whether the broader trend is up, down, or flat, and only favor entries aligned with that broader direction.
• A common construction compares a faster-moving average to a slower one: the faster average being above the slower one is read as supportive of an uptrend, and the reverse for a downtrend.

When It May Work

• Trend following tends to perform best in markets with sustained, persistent directional moves -- exactly the conditions where counter-trend approaches tend to struggle.
• It tends to underperform in sideways, mean-reverting conditions, where the 'trend' whipsaws back and forth, generating signals that reverse shortly after being triggered.

Main Risks

• Because trend indicators are built from past prices, they are inherently lagging -- confirmation of a new trend typically arrives after a meaningful part of the move has already occurred.
• A prolonged sideways market can produce a sequence of false signals in both directions, each with a small cost, that accumulate over time even though no individual signal was unreasonable given the information available at the time.

What the System Analyzes

• Fast and slow moving averages (e.g. EMA20 relative to EMA50)
• Price relative to those moving averages
• Consistency of the trend signal across the evaluation window

How It Works

• Rather than deeply analyzing every asset continuously (which is computationally expensive and often unnecessary), an opportunity-detection layer scores the full market universe using cheap, always-available signals, then applies deeper analysis only to the top-ranked candidates.
• The score blends several independent components -- for example acceleration of recent price movement, technical confirmation, and liquidity -- with soft penalties applied for conditions associated with lower-quality setups, such as an extreme, spike-like price move that is more likely to mean-revert than continue.

When It May Work

• This layered approach is most valuable when the opportunity set is large -- scanning dozens of assets for genuine outliers is impractical to do manually with consistent discipline.
• It is a discovery and ranking tool, not an execution decision on its own -- its output is a shortlist for further validation, not a standalone buy signal.

Main Risks

• A high rank reflects favorable readings on the specific signals measured -- it does not account for information the scoring model doesn't observe, such as breaking news or an exchange-specific liquidity event.
• Ranking systems can be noisy at the margin: an asset ranked 5th and one ranked 6th are not necessarily meaningfully different in quality.

What the System Analyzes

• Short- and medium-term price acceleration across the tracked universe
• Technical confirmation (trend/momentum alignment) for top-ranked candidates
• Liquidity (trading volume) and data-confidence scoring
• Risk flags for extreme, spike-like, or overextended price action

How It Works

• A risk engine sits between a strategy's signal and any actual order, and enforces limits that apply regardless of how confident the signal is: maximum notional per order, maximum number of simultaneous open positions, and instrument/asset eligibility.
• Critically, this layer is independent of the strategy layer -- it does not 'trust' the signal, it re-verifies every constraint itself on every single order, every time.

When It May Work

• Risk controls are not designed to improve returns -- they are designed to bound losses and ensure the system survives long enough for a genuine statistical edge, if one exists, to play out over many trades.
• Their value is most visible in hindsight, during periods of unusual volatility or strategy underperformance, when the limits prevent a bad stretch from becoming a catastrophic one.

Main Risks

• Risk controls can only bound the risks they are designed to measure -- they do not eliminate market risk, execution risk, or the possibility of loss.
• Overly conservative limits can meaningfully reduce a strategy's opportunity set; setting limits is itself a trade-off decision, not a free improvement.

What the System Analyzes

• Per-order notional against a fixed ceiling
• Current open-position count against a fixed maximum
• Asset/instrument eligibility for the current environment
• Whether the trading environment itself is correctly and safely configured before any order is considered

How It Works

• Position sizing converts a target risk (or target notional) into an actual order quantity, respecting the exchange's own precision requirements for that instrument -- rounding down, never up, so a sized order never exceeds its intended risk.
• A disciplined system separates 'is this a good idea' (the signal) from 'how much should this idea be sized' (a mechanical, rules-based calculation) -- the two are never conflated into a single subjective decision.

When It May Work

• Consistent, rules-based sizing is valuable in essentially all conditions -- its benefit isn't regime-dependent the way a specific strategy's edge can be. It is what allows a real edge, however small per trade, to compound reliably over many trades.
• It is particularly valuable during a losing streak, where consistent sizing prevents an emotional 'revenge sizing' response that has historically been one of the most damaging behavioral patterns in trading.

Main Risks

• Sizing decisions are still exposed to the underlying market risk of the position itself -- disciplined sizing bounds the damage from being wrong, it does not prevent being wrong.
• A sizing model calibrated to historical volatility can be caught off guard by a genuine volatility regime change that hasn't yet shown up in the recent data it's based on.

What the System Analyzes

• Target notional per new position
• The instrument's own minimum order size and size increment (read live from the exchange, never assumed)
• Current reference price at the moment of sizing

How It Works

• Beyond any single trade's own risk, an aggregate exposure limit caps the total capital committed across every simultaneously open position -- ensuring that even a sequence of individually reasonable decisions cannot, in combination, put an outsized share of capital at risk at once.
• This is typically enforced independently at two levels: a limit on the number of simultaneous positions, and a separate limit on their combined notional value, so neither constraint alone has to do all the work.

When It May Work

• This discipline matters most exactly when it's least intuitive to apply -- during a strong run where every recent signal has looked attractive, and the temptation to keep adding exposure is highest.
• It also protects against correlation risk: several positions that look diversified by name can move together in practice during a broad market event, and an aggregate cap limits the damage from that scenario even without predicting when it will occur.

Main Risks

• An aggregate limit is a blunt instrument -- it doesn't distinguish between ten genuinely uncorrelated positions and ten highly correlated ones; it caps total size regardless.
• Being capacity-constrained means a system may decline a new opportunity purely because existing exposure is already at its limit, even if that new opportunity looks attractive on its own merits.

What the System Analyzes

• Sum of notional value across all currently open, materially sized positions
• Current position count against a fixed maximum
• Whether adding a new position's target notional would breach the aggregate exposure ceiling