Skip to main content

mts1b-research — public API surface

Strategy discovery + ladder sweep + factor library + live executor + drift monitor.

Factor registry​

from mts1b_research.factors import register, get, FACTOR_REGISTRY

@register("f_my_momentum")
def f_my_momentum(panel, /, h: int = 21):
    return zscore_cross_sectional(panel.close[-1] / panel.close[-h] - 1)

# Discover
print(sorted(FACTOR_REGISTRY))
fn = get("f_my_momentum")

Strategy registry​

from mts1b_research.strategies import StrategyRegistry, StrategySpec

await StrategyRegistry.register(
    StrategySpec(
        strategy_id="momentum_v3",
        factor="f_momentum_12_1",
        params={"h_long": 252, "h_skip": 21},
        universe="us-large-cap",
        rebal="monthly",
        sizing="kelly_voltarget_12",
        cost_bps=5,
        enabled=True,
    )
)

# Query
active = await StrategyRegistry.active()
spec = await StrategyRegistry.get("momentum_v3")

# Lifecycle
await StrategyRegistry.disable("momentum_v3")
await StrategyRegistry.halve_allocation("momentum_v3", reason="early decay")
await StrategyRegistry.shadow("momentum_v3", reason="drift > -2.0σ")

Ladder sweep​

from mts1b_research.ladder import run_ladder

await run_ladder(
    factor_class="momentum",
    param_grid={
        "h_long": list(range(60, 365, 5)),
        "h_skip": [0, 5, 21, 42],
        "winsorize": [True, False],
    },
    universe="us-large-cap",
    start="2010-01-01", end="2024-01-01",
    cost_bps=5,

    # Stage configs (defaults shown)
    l1_keep_pct=0.01,
    l2_min_stability=0.5,
    l3_regimes=["bull", "chop", "crash"],
    l4_cost_multiplier=2.0,
    l5_max_candidates=5,
)

Writes survivors per stage to data/sweeps/ladder/<run_id>/.

Live executor​

from mts1b_research.strategies import live_executor

# Long-running worker
await live_executor.run(
    fund_id="paper-momentum",
    strategies_from="registry",   # or explicit list
)

What it does:

  1. At each rebalance window, looks up active strategies
  2. Computes factor on latest UniversePanel
  3. Applies sizer in mts1b-portfolio
  4. Diffs against current OMS positions
  5. Emits child orders via OMS

CLI: mts1b-research live-executor run --fund-id paper-momentum.

Dual-paper​

from mts1b_research.strategies import DualPaperExecutor

executor = DualPaperExecutor(
    strategy_id="momentum_v3",
    paper_fund="paper-momentum-demo",
    live_fund="live-momentum",
    diff_threshold_bps=10.0,
)
await executor.run()

Runs the SAME strategy code against both books; alerts on divergence > threshold.

Drift monitor​

from mts1b_research.ops import drift_monitor

# Polled (Prefect-scheduled)
result = await drift_monitor.measure(strategy_id="momentum_v3")
# DriftMeasurement(
#   strategy_id="momentum_v3",
#   backtest_ic=0.062, backtest_sharpe=1.43,
#   live_ic_30d=0.054, live_sharpe_30d=1.27,
#   drift_zscore=-0.24,
#   decay_status="NORMAL",   # or "WARN" | "HALVE" | "SHADOW"
# )

# Manual
await drift_monitor.evaluate_all()    # checks all active strategies

Auto-throttle thresholds:

drift_zscoreAction
< -0.5σTelegram warning
< -1.0σHalve sleeve allocation
< -2.0σShadow the strategy

Factory ensembles​

from mts1b_research.factory import canonical_alpha_factory, factory_ensemble

# Discover N candidate factors
candidates = await canonical_alpha_factory.discover(
    asset_class="equities",
    target_count=50,
    min_walkforward_sharpe=0.5,
    min_ic=0.02,
    min_stability=0.5,
)

# Compose into a meta-strategy
ensemble = factory_ensemble(
    candidates=candidates,
    method="hrp",                # or "equal" | "min_var"
)
# Returns a registered meta-strategy id

NATS subjects​

SubjectDirectionPayload
mts.v1.research.signals.publishedpublishSignal
mts.v1.research.ladder.stage_completepublishstage summary dict
mts.v1.research.drift.measuredpublishdrift snapshot dict
mts.v1.marketdata.bars.*subscribelatest bars for rebal
mts.v1.oms.fills.createdsubscribeobserve live execution

CLI​

mts mts1b-research factors list
mts mts1b-research strategy register --strategy-id X --factor Y --params '...'
mts mts1b-research strategy show --strategy-id X
mts mts1b-research strategies list --enabled-only
mts mts1b-research ladder run --factor-class momentum --param-grid '...'
mts mts1b-research drift show --strategy-id X
mts mts1b-research live-executor run --fund-id Y --strategies-from registry

See also​