The Strategy Analytics Tab,
Explained Module by Module
A practical guide to the advanced analytics panel inside AlphaNet's Strategy Details view — what each module shows, how to read the numbers, and why each one matters before you deploy capital into an AI trading strategy.
What this guide covers
Every strategy card in the AlphaNet AI Strategy Marketplace opens a Strategy Details panel with three tabs — Metrics, Analytics, and Trades. The Analytics tab is the newest of the three: instead of telling you what the strategy returned (Metrics) or which trades it placed (Trades), it tells you how the strategy behaves — which market environments it thrives in, how much pain it can inflict on the way, and how much directional risk it actually takes.
The Analytics tab contains three modules:
| Module | Question it answers | Used for |
|---|---|---|
| Market Regime Scoring | "In what kind of market does this strategy perform well or poorly?" | Deployment timing, variant selection, expectation setting |
| Drawdown Analysis | "How deep have the losses-from-peak been — recently and historically?" | Risk budgeting, position sizing, stress tolerance |
| Exposure Analysis | "How much net long / short risk does the strategy actually carry?" | Portfolio correlation control, leverage awareness, profile verification |
Throughout this guide we use Hackworth Prime – ZEC (Version V2) on the
PERP_ZEC_USDC market as the worked example, with the live values displayed on
27 July 2026. Every module also carries a small ⓘ info icon in its header inside the
app — hover over it in the platform for the built-in one-line definition.
Getting to the Analytics tab
https://trade.alphanet.global/strategies/?s=Hackworth+Prime&m=PERP_ZEC_USDC.
This is useful for sharing a strategy view — the Share button at the top of the panel
produces the same link.
Orientation: the Strategy Details header
Before the three modules, the Analytics tab inherits the metadata header of the Strategy Details panel. These fields frame everything below them, so read them first:
| Field | Example (Hackworth Prime – ZEC) | Meaning |
|---|---|---|
| Model / Engine | Hackworth Prime – ZEC | The strategy engine and the asset this instance trades. |
| Version | V2 | Engine generation. Newer versions incorporate risk-management, alpha-discovery, regime-detection and execution upgrades. |
| Tech Type | Quantitative AI (Deep Learning) | The modelling paradigm behind the signals. |
| Strategy Type | Multi-Alpha | An ensemble of many underlying alpha models rather than a single signal. |
| Market Profile | Neutral | The strategy's intended directional stance — Neutral, long-biased or short-biased. Check it against the Exposure Analysis module (Section 06). |
| Avg. Net Exposure | 27.15% | Average net directional exposure over the track record, as a share of allocated capital. |
| Strategy Capacity | Bar + POPULAR tag | How full the strategy is. When max capacity is reached, new capital cannot be deployed until total value deployed (TVD) falls. The bar fills from green toward amber/red as capacity saturates. |
Market Regime Scoring
What you see
The module is split into two side-by-side panels, each rating the strategy across three market states — Uptrend, Chop and Downtrend — on a 1 to 5 scale (1 = poor, 5 = excellent), shown as a five-segment bar plus a numeric score:
- Fast Regime (~Weekly Frequency) — performance in short-lived market states, the kind that play out over days to a week.
- Macro Regime (~Monthly Frequency) — performance in sustained, month-scale market environments.
How to read it
Each score answers: "When the market was in this state at this frequency, how good was this strategy at making (or protecting) money?" A 5.0 means the strategy historically excelled in that state; a 1.0 means that state is where it struggled most. Read the module comparatively — the spread between scores is more informative than any single value.
In the reference example:
- Fast panel (4.0 / 3.5 / 5.0): consistently strong across all weekly states — no score below 3.5 — and at its very best in fast downtrends. Short-biased and mean-reversion alphas inside the multi-alpha ensemble monetize sharp weekly sell-offs.
- Macro panel (5.0 / 1.0 / 2.5): far more polarized. The strategy is excellent in month-scale uptrends, passable in sustained downtrends, and weak in prolonged, directionless chop — the environment where trend-following signals get repeatedly faded and whipsawed.
Why this module matters
No alpha works in all weather. Deep-learning strategies in particular learn regime-conditional patterns, so their realized performance is always a function of which regimes actually occurred during the track record. Regime Scoring converts that hidden dependency into an explicit map. Under the hood, AlphaNet's engine stack classifies market state with a family of unsupervised regime models (trend, entropy/disorder, volatility expansion–contraction, volume, and cross-asset correlation families; the current engine generation tracks 27+ distinct regimes across horizons from 1H to 7D). The 1–5 scores you see are the condensed, user-facing summary of that machinery: per-regime quality ratings at weekly and monthly scale.
Practical uses
- Deployment timing. If your own read of the macro regime is "range-bound for months," a strategy scoring 1.0 on Macro Chop deserves a smaller allocation — or a wait.
- Variant selection. Engine families ship complementary profiles (e.g. the balanced Prime, the long-biased Trend, the short/mean-reversion-heavy OptimaShort). Regime scores are the objective basis for matching the variant to the environment.
- Building a "strategy of strategies." Combining strategies whose regime strengths are complementary (one excels in macro uptrends, another in chop) smooths the aggregate equity curve — the platform explicitly supports this multi-strategy construction.
- Expectation setting. A strategy with a 1.0 in the current regime will look "broken" for stretches. Knowing that in advance is what stops you from de-deploying at the exact bottom of its cycle.
Drawdown Analysis
What you see
Four tiles showing the strategy's maximum drawdown measured over four trailing windows: the last 30 days, 60 days, 90 days, and all time (the full track record, matching the Max Drawdown figure on the Metrics tab).
How to read it
A drawdown is the peak-to-trough decline of the equity curve — if the strategy's value rises to a high of 100 and then falls to 73 before recovering, that is a 27% drawdown. "Maximum drawdown within a window" is the deepest such episode whose peak and trough both sit inside that window. Reading the four tiles as a sequence gives you the term structure of pain:
- 30D 3.4% — recent trading has been calm: the worst give-back in the past month was shallow.
- 60D 21.3% — stretch the window one more month and a serious episode appears: within the last two months the strategy gave back roughly a fifth of its value from a peak.
- 90D 26.6% — the worst three-month episode is deeper still.
- All Time 26.6% — identical to the 90D figure, meaning the worst drawdown in the entire 2.5-year track record happened within the last 90 days.
Why this module matters
A single lifetime max-drawdown number hides when losses happen. The windowed view separates three very different risk questions:
- "What's normal right now?" (30D/60D) — sets expectations for the routine give-backs you must psychologically and financially tolerate while deployed.
- "What's the worst I should be prepared for?" (All Time) — the stress number for sizing. If 26.6% peak-to-trough would force you to capitulate, the allocation is too large, full stop.
- "Is risk stable or escalating?" (the pattern across windows) — monotone growth from 30D → All Time is healthy. A 30D tile that suddenly matches the All-Time tile means the worst episode ever is happening now.
This directly feeds the risk budget: AlphaNet's engine internally sizes positions against a remaining-drawdown budget, and the platform's own capacity logic responds to volatility and liquidity conditions — the windowed drawdown view is the user-facing counterpart you can monitor from outside.
Exposure Analysis
What you see
A time-series chart of the strategy's net directional exposure over its history, oscillating between three zones — Long (above zero, green), Neutral (at the zero line), and Short (below zero, pink) — plus three summary tiles: Current Net Exposure, Max Net Long Exposure, and Max Net Short Exposure.
How to read it
The vertical axis is net exposure expressed as a fraction of allocated capital: +0.40 means the strategy is net long 40% of its capital; −0.20 means net short 20%; 0 means flat/market-neutral. "Net" means longs and shorts offset each other — only the residual directional bet is shown.
The three tiles compress the whole chart into a risk envelope:
- Current Net Exposure +0.40 — right now the strategy sits at its historical maximum long posture.
- Max Net Long +0.40 / Max Net Short −0.20 — the band is asymmetric 2:1 toward the long side: the engine is willing to commit twice as much net capital upward as downward, even though its stated Market Profile is "Neutral."
Why this module matters
- Verify the strategy is what it claims to be. A "neutral" strategy whose chart pins at +0.9 for months is quietly a long-biased bet on the asset. Here, the bounded ±band and the frequent returns to the zero line are exactly what a genuinely tactical, multi-alpha neutral engine should look like.
- Control portfolio-level correlation. Your true market risk is the sum of net exposures across everything you run. If three deployed strategies all sit at +0.40 simultaneously, you are effectively long the crypto complex with leverage even though each looks "neutral" alone. This module is where you audit that overlap.
- See the dynamic sizing engine at work. Modern AlphaNet engines size positions as a function of signal strength, the prevailing volatility regime, and remaining risk budget — so a healthy chart should be variable: expanding into high-conviction directional regimes, contracting toward zero in uncertainty. A flat-line chart would signal a static, less adaptive strategy.
- Anticipate drawdown drivers. Read together with Drawdown Analysis: deep drawdowns that coincide with extreme exposure readings tell you losses came from committed directional bets; drawdowns at near-zero exposure point to whipsaw on the tactical book instead.
Using the modules together
The three modules are designed to be read as one instrument. A practical pre-deployment checklist:
After deployment: a monitoring cadence
| Frequency | Watch | Action trigger |
|---|---|---|
| Weekly | 30D drawdown tile vs. your tolerance; Current Net Exposure vs. the rest of your book | 30D drawdown approaching the All-Time worst, or aggregate net exposure drifting one-sided |
| Monthly | Regime scores for migration; 60D/90D drawdown pattern; ROI-By-Month on the Metrics tab | Regime profile shifting against the strategy's weak scores; worst-case window getting more recent |
| On regime change | Re-run the full Section 07 checklist when the macro state flips (e.g. uptrend → chop) | Consider re-balancing across strategy variants whose scores complement the new state |
How Analytics complements the other two tabs
- Metrics tab gives the verdict: ROI (1869.08%), period (2y 6mo 26d), Max Drawdown (26.61%), Win Rate (48.93%), trade counts, average winner/loser (+3.83% / −2.21% — the asymmetric payoff profile typical of trend-capture systems that win less than half the time but win big), Sharpe (2.69), plus the ROI Curve and ROI-By-Month charts. Analytics explains the conditions and posture behind those numbers.
- Trades tab gives the evidence: individual entries and exits. Use it to validate what Analytics implies — e.g. confirm that high-exposure windows line up with the regime states the strategy scores well in.
Quick-reference tables
Every element of the Analytics tab
| Element | Format | Definition | Healthy sign | Warning sign |
|---|---|---|---|---|
| Fast Regime scores | 3 × (1–5) | Strategy quality in weekly Uptrend / Chop / Downtrend states | No score near 1 in the current weekly state | Low score in the state you expect next |
| Macro Regime scores | 3 × (1–5) | Strategy quality in month-scale Uptrend / Chop / Downtrend | High score in the prevailing macro regime | Prevailing macro regime matches the strategy's weakest score |
| Drawdown 30D / 60D / 90D | % × 3 | Deepest peak-to-trough decline within each trailing window | Shallow recent tiles relative to All Time | Recent tiles converging on or exceeding All Time |
| Drawdown All Time | % | Deepest drawdown of the full track record (= Metrics tab figure) | Within your personal tolerance with margin | Larger than what would make you capitulate |
| Exposure chart | time series | Net directional exposure over time, in fractions of allocated capital | Variable path consistent with stated Market Profile | Persistent one-sided pinning that contradicts the profile |
| Current Net Exposure | ± fraction | Net long/short posture right now (0.40 ≈ 40% of capital net long) | Inside the historical band | At band extreme while your portfolio is already same-direction |
| Max Net Long / Short | ± fraction | Historical extremes of the net exposure band | Symmetric-ish band matching the strategy's mandate | Band far wider than the stated profile implies |
Worked example values (Hackworth Prime – ZEC · V2 · 27 Jul 2026)
| Module | Reading | One-line interpretation |
|---|---|---|
| Fast Regime | Up 4.0 · Chop 3.5 · Down 5.0 | Strong in all weekly states; exceptional in fast downtrends. |
| Macro Regime | Up 5.0 · Chop 1.0 · Down 2.5 | Superb in sustained uptrends; prolonged flat chop is its weakness. |
| Drawdown | 30D 3.4% · 60D 21.3% · 90D 26.6% · All 26.6% | Worst-ever episode occurred within the last 90 days; recent month calm. |
| Exposure | Now +0.40 · Max long +0.40 · Max short −0.20 | Tactical band with a 2:1 long-side asymmetry; currently at max long. |
| Context | Avg net exposure 27.15% · Neutral profile | Design stance confirmed by the exposure band and its average. |
Limitations & caveats
- Everything here is backward-looking. Regime scores, drawdowns and exposure bands are computed from the strategy's realized track record. They describe where the engine has been strong, not where it will be.
- Proprietary internals. The exact metric behind the 1–5 regime scale, the regime classifier's features, and the exposure sampling frequency are not disclosed on the page. Interpret values comparatively (across regimes, across strategies, across time), not as absolute measurements.
- Window statistics hide path detail. Drawdown tiles show depth, not frequency, duration, or recovery speed; exposure tiles show extremes, not typical day-to-day variation.
- Version sensitivity. Values are engine-version specific. The reference strategy runs Hackworth V2; the newer V3 generation (with hard stop-loss overlays, expanded regime detection and upgraded execution) may print materially different analytics on its markets.
- Capacity and fees are outside these modules. The Analytics tab does not reflect capacity exhaustion, funding costs, or execution slippage — those live in the header, the FAQ, and realized Metrics.