Monstra loading art

Strategy Platform

Loading Monstra

Preparing the next route...

MonstraTM
  • Home
  • About
  • Monstra Bots
  • Vectors
  • Titans
  • Armor
  • Heartbeat
MonstraTM

Monstra is a simulation and educational platform. Bot signals are algorithmic outputs, not personalized advice. Users make their own decisions, and past simulated performance does not guarantee future results.

Product

  • Brokerage
  • Bots

Explore

  • About
  • Blog
  • Support

Legal

  • Terms of Service
  • Privacy Policy
  • Risk Disclosure
  • Brokerage Disclosure

© 2026 Monstra. All rights reserved.

Built for research, simulation, and strategy exploration on Monstra's dark-first platform.

Operational intelligence

Monstra Heartbeat

Monstra runs live trading algorithms end to end: market data in, vectors computed, a tradable universe built, algorithms scored, signals produced, and performance measured — all shown here as it actually runs.

This page follows one path through the system, left to right: market data, vector computation, universe construction, algorithm evaluation, signal, and measurement. Every number here is either queried live or explicitly marked as not yet instrumented — nothing is invented. Click a stage below to jump to it.

Last write
1d ago
Across vectors, signals, and worker runs
Vector writes, 24h
0 writes in the last 24h
⚠Market datastale
Sep 11, 6:05 PM
⚠Vector computationstale
Sep 11, 9:21 PM
⚠Universe constructionstale
Aug 14, 7:28 PM
✓Algorithm evaluationfresh
Sep 11, 6:37 PM
⚠Signalstale
Sep 11, 6:16 PM
⚠Measurementstale
Sep 11, 12:00 AM
Current wave progress meter: not yet instrumented on the public data layer — the vector refresh scheduler tracks per-ticker cadence internally, but doesn't yet expose an in-flight wave percentage. Reported rather than faked.
Market data

The first stage: prices in

Everything downstream depends on this landing first. This is the same real timestamp the "Market data" pip in the strip above tracks — not a separate query.

Latest close ingested
Sep 11, 6:05 PM
Latest close ingested
Per-vendor coverage (which data provider supplied which symbols, and where any are missing) is not yet broken out as its own metric on the public data layer — reported as a gap rather than estimated or invented.
Vectors

Live vector refresh output

This feed reflects the real trading.Vector refresh wave, so this section shows the market-data pipeline actually computing features.

Coverage
96.8%
6,903 of 7,129 live-safe tickers
Updated 24h
0
Rows touched in the last day
Tracked rows
6,905
Current rows in trading.Vector
Missing
226
Live-safe symbols still not covered
Refresh completion
Latest write Sep 11, 9:21 PM

What each vector actually means

Six decimals per stock, defined precisely rather than left as bare numbers.

Momentum

Absolute price momentum across multiple horizons — higher means the stock has been going up, on its own terms, not relative to the market.

Formula
Weighted blend of 1m/3m/6m/1y returns (20/20/15/10%) plus momentum acceleration (1m-vs-3m, 3m-vs-6m) and 3-month trend slope, each min-max normalized to a fixed range (e.g. 1m return clamped to [-30%, +30%]).
Universe

From listed symbols to today's holdings

This funnel runs at two different scales, shown as two blocks below. The top block starts from a manually-rebuilt research-universe snapshot (as of 2026-08-13: 200 symbols listed) — a different, and currently much smaller, pool than the 7,129 live-safe tickers and 6,903vector-covered tickers the Vectors section above tracks continuously. Past the "scale shift" divider, the bottom block drops to a third scale: only the five showcased bots' own curated universes. These numbers are not meant to reconcile with each other — they measure different pools by design.

Each row is what survived the previous one. A row marked "No change" shows the same count as the row above it, with its reason shown inline rather than hidden behind a click. Rows with sample tickers can still be expanded for the exact rule and what it dropped.

Symbols listed
200
Candidate pool: current Nasdaq/NYSE listings, common equity only
As of 2026-08-13
Pass liquidity gate
33
Exchange, security type, price ≥ $5, market cap ≥ $1B, 20d avg dollar volume ≥ $10M, ≥504 days of history
167 dropped
AAA
Algorithms

Five strategies, side by side

Same axes on every card — factor exposure is the weighted average of each bot's own current holdings' vector scores, not a declared target.

Bellitus (Force)

Momentum rotation within a fixed, theme-based universe — ranks names by trailing momentum and holds the top tier, with a benchmark filter and kill switch.

Momentum
0.68
Rel. strength
0.73
Stability
0.37
Aptet fit
n/a
Force fit
n/a
Draco fit
n/a
Signals

Latest signals by bot

Every row stays traceable by trading date, session, and run id.

Bellitus (Force)
2026-09-11 / afternoon / 20260911-PM
unknown
Performance window: No history · No History
VREX 40.00%ZBRA 30.00%TAYD 20.00%TTEK 10.00%
Bellitus (Aptet)
2026-09-09 / afternoon / 20260909-PM
unknown
Performance window: No history · No History
ASML 25.00%TWIN 25.00%VREX 25.00%ZBRA 25.00%
Bellitus (Draco)
2026-09-11 / afternoon / 20260911-PM
unknown
Measurement

Does it work, and how would we know

Each bot's equity, indexed to 100 at t0, against VOO on the same axis. Drawdown shares the x-axis beneath. Sample-size caveats are shown, not buried.

Bellitus (Force)
Not enough equity history to chart.
Since
n/a
n trading days
0
Net return
n/a
Sharpe (ann.)
n/a
Max drawdown
n/a
Source
unknown
n=0trading days — short of the ~252-day (one trading year) floor commonly used before a Sharpe ratio is treated as statistically meaningful. Read these numbers as directional, not conclusive.
Bellitus (Aptet)
Not enough equity history to chart.
Honesty panel

What this page doesn't claim

The thesis of this page is that an honest gap reads as competence, not as a failure. Everything below is real, current, and unresolved as of this writing.

Simulated vs. live

Each bot's Measurement card carries its own source label (Live Trading, Paper, Backfill, or Mixed), taken directly from the stored performance source on each equity row — not inferred. A backfilled segment is a historical simulation, not a live track record.

Point-in-time vs. restated

The research universe (Universe construction, above) uses current-day listings and current fundamentals, not historical constituents. It is not survivorship-bias-free and there is no enforced lag between data availability and its use in a signal. This is stated plainly there, not buried here.

What isn't built yet
  • Current wave progress meter — the vector refresh scheduler tracks per-ticker cadence but doesn't expose an in-flight percentage.
  • Universe churn — the snapshot table this needs doesn't exist in this database yet; nothing has been frozen to diff.
  • Gross-vs-net-of-costs on live segments — live equity already reflects real fills, but isn't separately decomposed into gross and assumed-cost components.
  • A true "delisted" exclusion reason — the closest real signal is unsupported-exchange/security-type, which is a proxy, not a direct delisting flag.
What would be fixed first

Scheduling the universe rebuild and snapshot-freeze on a cadence, instead of by hand, would fix two things at once: the funnel and coverage numbers above would stop lagging (currently weeks old), and universe churn would go from unpopulated to real. That's the single highest-leverage gap on this page.

Every number on this page is either queried live, derived from a query already shown, or explicitly marked as not yet instrumented. Nothing here is interpolated or invented.
Range
0-1, fixed-range normalized. 0 = weakest, 1 = strongest. Not cross-sectional — a single ticker's score is meaningful in isolation.
Cadence
Recomputed on the base vector refresh wave (daily-ish, see the write-volume sparkline above).
Relative strength

Performance versus the benchmark, not versus zero. A stock up 8% while the benchmark is up 12% scores as weak; a stock down 3% while the benchmark is down 10% scores as strong.

Formula
Weighted blend of excess return (stock return minus benchmark return) over 1m/3m/6m/1y (20/25/25/15%) plus relative-strength slope and acceleration, min-max normalized.
Range
0-1, fixed-range normalized. 0 = leading the benchmark down or lagging it up, 1 = clear outperformance.
Cadence
Recomputed on the base vector refresh wave.
Stability

Risk stability — higher means safer: less volatile, smaller drawdowns, thinner tail risk. Oriented so high is always good, unlike raw volatility.

Formula
Weighted blend of inverted 1y/2y volatility, inverted downside volatility, current and max 1y drawdown, 1y trend R², and inverted 95% expected shortfall.
Range
0-1, fixed-range normalized. 0 = high volatility and deep drawdowns, 1 = calm and shallow drawdowns.
Cadence
Recomputed on the base vector refresh wave.
Aptet fit

How well the Aptet engine has historically performed on this specific ticker — a backtested, confidence-adjusted fit score, not a live signal.

Formula
Grid search over Aptet's holding-count and adaptation-speed parameters, backtested on forward returns (1/5/10 trading days) since 2025-01-01. The best combo's quality is penalized by parameter-neighbor stability and sample confidence (thin history or unstable neighboring parameters lower the score even if the single best combo looked good).
Range
0-1. Labeled poor (<0.35), weak (<0.50), moderate (<0.65), good (<0.80), strong (≥0.80).
Cadence
Recomputed on a tiered cadence (weekly/biweekly/monthly/quarterly per ticker) — much slower than the base vector, and can lag it by days.
Force fit

How well the Force engine (momentum rotation across a fixed sector universe) has historically performed on this specific ticker.

Formula
Same grid-search-plus-confidence-penalty methodology as Aptet fit, run against Force's own parameter grid and forward-return windows.
Range
0-1. Labeled poor (<0.35), weak (<0.50), moderate (<0.65), good (<0.80), strong (≥0.80).
Cadence
Recomputed on a tiered cadence (weekly/biweekly/monthly/quarterly per ticker).
Draco fit

How well the Draco engine (stop-loss-governed, regime-aware) has historically performed on this specific ticker.

Formula
Same grid-search-plus-confidence-penalty methodology as Aptet/Force fit, run against Draco's own parameter grid (holding-period cap, stop-loss percent, re-entry cooldown) and forward-return windows.
Range
0-1. Labeled poor (<0.35), weak (<0.50), moderate (<0.65), good (<0.80), strong (≥0.80).
Cadence
Recomputed on a tiered cadence (weekly/biweekly/monthly/quarterly per ticker).

Coverage, by cause of absence

Why the research universe's 200 candidates shrink to a covered set, as of 2026-08-13.

Security type excluded92 (55.1%)
Unsupported exchange29 (17.4%)
Below liquidity/quality floor43 (25.7%)
Vendor data gap3 (1.8%)

Vector age across the universe

Age of the newest vector per symbol — proving the heartbeat reaches past the most liquid names.

0
6,880
0
0
25
< 1 day
1-3 days
3-7 days
7-14 days
> 14 days

Live write feed

Polls for fresh rows every few seconds and merges them in, capped at 50. There is no change-log table, so this shows the current value at each write, not an old-to-new delta.

Last write 1d ago50 rows shown across 2 write batches, capped at 50
9:21:38 PM27 rows writtenshow rowshide rows
SymbolFactorValue
ZYMEmomentum0.672
ZJANmomentum0.511
ZSTKmomentum0.602
ZROZmomentum0.454
ZMmomentum0.463
ZIMmomentum0.770
ZKHmomentum0.511
ZSLmomentum0.444
ZWSmomentum0.392
ZTSmomentum0.393
ZTRmomentum0.506
ZNTLmomentum0.454
ZIPmomentum0.387
ZIONmomentum0.518
ZIONPmomentum0.463
ZJULmomentum0.510
ZNOVmomentum0.515
ZSEPmomentum0.513
ZUMZmomentum0.342
ZVRAmomentum0.555
ZSmomentum0.526
ZVOLmomentum0.620
ZVIAmomentum0.361
ZURAmomentum0.626
ZTOmomentum0.396
ZOCTmomentum0.517
ZLABmomentum0.596
9:21:33 PM23 rows writtenshow rowshide rows
SymbolFactorValue
XRAYmomentum0.408
XQQImomentum0.516
XRMImomentum0.519
XRNmomentum0.545
XPROmomentum0.539
XVVmomentum0.533
XYLGmomentum0.526
XPHmomentum0.568
XPEVmomentum0.244
XPOmomentum0.392
XPAYmomentum0.518
XPONmomentum0.587
XRPmomentum0.755
XTENmomentum0.478
YALAmomentum0.486
XPELmomentum0.559
XONEmomentum0.506
XPOFmomentum0.194
XOMmomentum0.615
XOVRmomentum0.543
XPERmomentum0.301
XPmomentum0.817
XYZGmomentum0.548

Throughput

Rows/sec, last hour
0.00
Writes, last hour
0
Fit-score refresh cadence
biweekly14
monthly13,414
quarterly5
weekly160

Stock focus — one vector stock at a time

Show stocks updated just nowShow latest tracked stocksOpen vectors
SMCI
Information Technology
Fit analysis ready
Base vector Sep 11, 9:17 PM

SMCI currently leans Draco with momentum 0.74, relative strength 0.64, and stability 0.12.

Fitness pass Sep 9, 2:58 AM
Momentum
0.74
Relative strength
0.64
Stability
0.12
Aptet fit
0.16
Force fit
0.12
Draco fit
0.71
ZIM *XOM *WMT *WMB *WELL *VST *VSAT *VRTX *VLO *V *USB *URI *UNP *UNH *UMC *UEC *UBER *TXT *TT *TSN *TSM *TSLA *TRGP *TMUS *TMO *TDG *SYK *SU *STEM *SQM *SO *SNOW *SMCI *SLB *SHG *SCHW *SBLK *RTX *ROL *RIOT ** has completed Aptet/Force/Draco fit analysis
2026-09-11
6,880
2026-07-29
25
etf_excluded
AAAAunsupported_exchange
AAACetf_excluded
AAADetf_excluded
AAAPetf_excluded
AAAUunsupported_exchange
AACBmarket_cap_below_minimum
AACBRright_excluded
Pass vector-coverage gate
No change33
Vector v2 contract match and feature coverage ≥ 90%
0 dropped — gate cleared by every stage-A survivor in this snapshot
Pass live-safe gate
No change33
Broker-tradable: tradable, not an ADR, Monstra live-safe flag set (app.alpaca_assets)
0 dropped — every eligible ticker in this snapshot is also live-safe
In tradable universe
No change33
Has a current row in trading.Vector
0 dropped — every live-safe eligible ticker already has a vector
Scale shift: five showcased bots, not the market-wide universe
Scored by the showcased algorithms
94
Union of the five showcased bots' configured tradable universes
Scale shift: from here down, only the five showcased bots, not the market-wide universe above
Held today
8
Scored but not currently held, per the latest signal session
86 scored but not held
AAnot in today's holdingsAAPLnot in today's holdingsACAnot in today's holdingsADBEnot in today's holdingsAEMnot in today's holdingsAIRnot in today's holdingsAMDnot in today's holdingsAMZNnot in today's holdings
Point-in-time correctness

As-of date: 2026-08-13. Membership is only built when the pipeline is run by hand — it is not rebuilt on a schedule, so this date can lag today by weeks.

Survivorship bias: not handled. The candidate source is today's Nasdaq/NYSE listing files, which only reflect currently-listed tickers. A name delisted since the as-of date is silently absent from history — this universe is not survivorship-bias-free.

Look-ahead: market cap and other metadata come from the data vendor's current values at build time, not the values as they stood on the as-of date. There is no enforced lag between data availability and its use in a signal.

Fundamentals: restated/current, not point-in-time. A future point-in-time constituent database would be required to fix this — acknowledged gap, not yet built.

The rules that define the tradable set

Same data as the funnel above, with the actual rule text, threshold, and why it exists.

  1. 1
    Exchange in the supported set (NMS, NYSE, NYSE American, NYSE Arca, and other Nasdaq-file-covered listings)
    Excludes OTC/pink-sheet names the broker won't reliably fill.
    29 dropped
  2. 2
    Common equity only — no ETF/ETP/ETN, fund, preferred, warrant, right, unit, ADR, closed-end fund, or SPAC
    Keeps the universe to single-company stocks the vector factors are designed for.
    92 dropped
  3. 3
    Price ≥ $5.00
    Sub-$5 names carry outsized slippage and manipulation risk relative to position size.
    22 dropped
  4. 4
    Market cap ≥ $1B
    Below this, index/fund flows dominate the name more than company fundamentals.
    19 dropped
  5. 5
    20-day average dollar volume ≥ $10M
    The floor for entering/exiting a position without materially moving the price.
    4 dropped
  6. 6
    ≥ 504 trading days (~2 years) of price history
    The vector's longest lookback windows (1y, momentum acceleration) need this much history to mean anything.
    1 dropped
  7. 7
    Vector v2 contract match and feature coverage ≥ 90%
    A ticker with too many missing features produces an unreliable, mostly-null vector.
    0 dropped

Daily universe churn

Symbols entering and leaving the eligible set, day over day.

2026-08-13
-134 / +18
AdditionsRemovals
Top position
40.0%
HHI
0.300
Turnover/rebalance
2.5
Mean hold
4d
Current holdings
VREX 40.0%ZBRA 30.0%TAYD 20.0%TTEK 10.0%
Bellitus (Aptet)

Adaptive holding-count strategy — widens or narrows how many names it holds based on regime, rather than a fixed top-N.

Momentum
0.64
Rel. strength
0.74
Stability
0.47
Aptet fit
0.62
Force fit
0.65
Draco fit
0.11
Top position
25.0%
HHI
0.250
Turnover/rebalance
1.6
Mean hold
3d
Current holdings
ASML 25.0%TWIN 25.0%VREX 25.0%ZBRA 25.0%
Bellitus (Draco)

Regime-aware, stop-loss-governed strategy — 10% stop loss, exits any position after 126 trading days regardless of ranking.

Momentum
0.57
Rel. strength
0.56
Stability
0.50
Aptet fit
n/a
Force fit
n/a
Draco fit
n/a
Top position
33.3%
HHI
0.333
Turnover/rebalance
1.5
Mean hold
11d
Current holdings
VMI 32.0%TRMB 32.0%VREX 32.0%
Cuprum (Alpha2)

Sector rotation — ranks six sector proxies by 20-day momentum, then buys rank-weighted stocks within the triggered sector(s), rebalanced every 5 days.

Momentum
n/a
Rel. strength
n/a
Stability
n/a
Aptet fit
n/a
Force fit
n/a
Draco fit
n/a
Top position
0.0%
HHI
0.000
Turnover/rebalance
1.3
Mean hold
3d
Current holdings
No holdings in the latest signal.
Dux (Echo1)

Lead-lag pair trading — discovers which stocks' moves predict others', then allocates to followers whose leaders are currently signaling, under a market-trend filter and drawdown guard.

Momentum
n/a
Rel. strength
n/a
Stability
n/a
Aptet fit
n/a
Force fit
n/a
Draco fit
n/a
Top position
0.0%
HHI
0.000
Turnover/rebalance
2.7
Mean hold
4d
Current holdings
No holdings in the latest signal.

Return correlation

Pairwise correlation of daily returns over the trailing three months. Published as-is, whatever it shows.

Bellitus (Force)
Bellitus (Aptet)
Bellitus (Draco)
Cuprum (Alpha2)
Dux (Echo1)
Bellitus (Force)1.00n/an/an/an/a
Bellitus (Aptet)n/a1.00n/an/an/a
Bellitus (Draco)n/an/a1.00n/an/a
Cuprum (Alpha2)n/an/an/a1.00n/a
Dux (Echo1)n/an/an/an/a1.00

Bot creation and update feed

Grounded in real strategy rows for the five showcased bots.

Bellitus (Aptet)
aptet · owner official
Created Jul 30, 4:49 PM · Updated Sep 9, 6:26 PM
Cuprum (Alpha2)
alpha2 · owner official
Created May 21, 5:34 AM · Updated Aug 27, 4:32 PM
Dux (Echo1)
echo1 · owner official
Created Jun 18, 7:44 AM · Updated Aug 1, 12:24 AM
Bellitus (Force)
force · owner official
Created May 19, 5:36 PM · Updated Aug 1, 12:23 AM
Bellitus (Draco)
draco · owner official
Created Jul 30, 4:49 PM · Updated Jul 30, 4:49 PM
Performance window: No history · No History
VMI 32.00%TRMB 32.00%VREX 32.00%
Cuprum (Alpha2)
2026-08-27 / morning / 20260827-AM
unknown
Performance window: No history · No History
No holdings parsed from the latest signal row.
Dux (Echo1)
2026-09-11 / afternoon / 20260911-PM
unknown
Performance window: No history · No History
No holdings parsed from the latest signal row.
Since
n/a
n trading days
0
Net return
n/a
Sharpe (ann.)
n/a
Max drawdown
n/a
Source
unknown
n=0trading days — short of the ~252-day (one trading year) floor commonly used before a Sharpe ratio is treated as statistically meaningful. Read these numbers as directional, not conclusive.
Bellitus (Draco)
Not enough equity history to chart.
Since
n/a
n trading days
0
Net return
n/a
Sharpe (ann.)
n/a
Max drawdown
n/a
Source
unknown
n=0trading days — short of the ~252-day (one trading year) floor commonly used before a Sharpe ratio is treated as statistically meaningful. Read these numbers as directional, not conclusive.
Cuprum (Alpha2)
Not enough equity history to chart.
Since
n/a
n trading days
0
Net return
n/a
Sharpe (ann.)
n/a
Max drawdown
n/a
Source
unknown
n=0trading days — short of the ~252-day (one trading year) floor commonly used before a Sharpe ratio is treated as statistically meaningful. Read these numbers as directional, not conclusive.
Dux (Echo1)
Not enough equity history to chart.
Since
n/a
n trading days
0
Net return
n/a
Sharpe (ann.)
n/a
Max drawdown
n/a
Source
unknown
n=0trading days — short of the ~252-day (one trading year) floor commonly used before a Sharpe ratio is treated as statistically meaningful. Read these numbers as directional, not conclusive.
Worker runs, last 30 days: 65.8% completed.