Fix uneven multi-battery charging from deadband concentration (#523)

[tomquist]

Fixes #523. The share_imbalance_w metric this relies on landed in #525 (merged), so this PR targets develop directly and the CI steering-eval comparison shows the balance change against a base that already has the metric.

Root cause

Deadband concentration (concentrate_deadband, added in 2.2.0 via #480) hands the whole small grid correction to the single most-active battery near steady state so it clears the firmware input deadband, and bypasses balance correction for that tick. But it fired whenever the grid error was small regardless of how the pool was split. An out-of-balance pair (two Venus E3 at ~88 W and ~890 W) produces no grid error — the sum is correct — so the only thing that could equalize them is balance correction, which concentration was suppressing. Result: the imbalance was pinned forever. This is absent in 2.1.2, which is why rolling back fixed it.

Fix (two parts)

1. Gate concentration on the pool already being balanced — skip it (and let balance correction equalize) whenever any battery is at least balance_deadband off its weight-proportional fair share.
2. Make equalization grid-neutral — the grid-direction sign clamp used to zero any residual opposing the grid. fair_share always carries the grid's sign, so that clamp only ever fired on the balance-correction term — near steady state, exactly the over-served battery's "back off" half of the swap. That left equalization one-sided (only the under-served battery moved), pushing the pool's net output around and disturbing the grid. Now only the grid-tracking half is clamped; the balance redistribution (≈zero-sum across the phase) goes through, so the over-served battery backs off by exactly what the under-served one takes on. Inert during load steps; only engages near steady state.

Mirrored on the C++/ESPHome side per the parity rule. New regression tests cover both the gate and the two-sided (grid-neutral) swap.

Measured effect (CI steering-eval, seed-averaged across 31 scenarios)

Metric	Base	Head	Δ
share_imbalance_w	157.7	63.7	−60% ✅
overshoot_max_w (guardrail)	91.4	86.7	−5% ✅
overshoot_mean_w	51.0	44.4	−13% ✅
settle_mean_s	30.9	36.1	+17%
grid_p2p_w (guardrail)	191.0	203.1	+6% ⚠️
battery_travel_w_per_h	19619	21414	+9%
cost_regret_ct / grid_rms / mean_abs	~flat	~flat	+0–4%

Priority-weighted verdict: −1.4% (better) — vs +3.2% worse with the gate alone (which tripped overshoot_max_w +24%). The grid-neutral swap turned the fix from a net regression into a net improvement: #523's imbalance is fixed (−60%, per-scenario −22% to −86%), overshoot now improves, and what remains is the inherent cost of equalizing two batteries (you have to move power between them): battery_travel +9%, settle +17%, and grid_p2p_w +6% (a smaller guardrail flag, traded down from the +24% overshoot one). All confined to multi-battery same-phase setups; single-battery scenarios are untouched.

This effectively restores 2.1.2-style balancing (what the reporter rolled back to and prefers) without 2.1.2's overshoot.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved battery balancing so multiple batteries on the same phase stay more evenly split during steady-state charging and discharging.
  • Tightened near-zero correction behavior to avoid hiding real imbalances and reduce one-sided “sticking” in edge cases.
  • Made concentration behavior safer by applying it only when the battery pool is already balanced.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 4527555e-c9eb-45fd-897a-7c433ab19c04

📥 Commits

Reviewing files that changed from the base of the PR and between c0c41ac and 1e3bc2d.

📒 Files selected for processing (5)

• CHANGELOG.md
• esphome/components/ct002/balancer.cpp
• esphome/components/ct002/balancer.h
• src/astrameter/ct002/balancer.py
• tests/test_balancer.py

---

Walkthrough

CT002 balancer now requires a concentration pool to be balanced before deadband concentration runs, and it preserves balance-correction residuals when grid-tracking signs disagree. The Python mirror, tests, and changelog were updated to match.

Changes

CT002 deadband balance fix

Layer / File(s)	Summary
Concentration gate and helper esphome/components/ct002/balancer.h, esphome/components/ct002/balancer.cpp, src/astrameter/ct002/balancer.py, CHANGELOG.md	Adds a pool-balance check for concentration, wires it into the deadband gate in both implementations, and records the imbalance regression in the changelog.
Residual split esphome/components/ct002/balancer.cpp, src/astrameter/ct002/balancer.py	Changes residual handling so sign disagreement clamps only the grid-tracking part while keeping the balance-correction portion intact.
Deadband concentration tests tests/test_balancer.py	Updates concentration and equalization coverage for balanced pools, imbalanced pools, and zero-weight participant behavior.

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~45 minutes

Possibly related PRs

• tomquist/AstraMeter#480: Updates the same CT002 deadband concentration and residual path in the allocation logic.
• tomquist/AstraMeter#429: Also changes CT002 residual/sign handling in balancer.cpp.
• tomquist/AstraMeter#475: Modifies the CT002 residual computation path with additional damping after clamping.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title is concise and accurately describes the main fix to uneven multi-battery charging.
Linked Issues check	✅ Passed	The changes gate concentration on balanced pools and preserve balance correction, matching #523's goal to restore even charging/discharging.
Out of Scope Changes check	✅ Passed	The changelog, code, and tests all support the balancing fix; no unrelated changes are indicated.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/stoic-planck-2r133k

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Steering evaluation (base vs head)

Overall: 4 improved, 11 regressed, 0 unchanged across 15 metrics — mean +3.9% (worse).

Priority: priority-weighted -1.7% (better) — ⚠️ do-no-harm guardrail regressed: grid_p2p_w +5%.

Lower is better for every metric. See src/astrameter/simulator/evaluation.py for definitions.

Metrics are the per-scenario mean of 5 seeds.

Aggregate — mean across 31 scenarios

Metric	Base	Head	Δ
settle_mean_s	30.6	36.3	+19%
settle_p95_s	49.6	63.5	+28%
unsettled_events	0.3	0.5	+67%
overshoot_mean_w	51.6	45.4	-12%
overshoot_max_w	105.0	97.7	-7%
band_crossings_per_h	306.4	306.8	+0%
grid_p2p_w	251.1	263.9	+5%
grid_rms_w	246.4	246.3	-0%
steady_rms_w	113.9	114.3	+0%
mean_abs_grid_w	96.8	97.8	+1%
share_imbalance_w	157.2	62.7	-60%
avoidable_import_wh	35.0	36.2	+3%
avoidable_export_wh	25.9	26.2	+1%
cost_regret_ct	0.81	0.85	+5%
battery_travel_w_per_h	21636.2	23435.9	+8%

📊 Interactive grid-power charts (zoom / hover / toggle series) are in the self-contained steering-eval-report.html report — see the link below (it opens directly in the browser).

What do these metrics mean?

Metric	Meaning
settle_mean_s	Mean seconds after a load/PV step for grid power to return inside the ±25 W settle band and hold for 10 s (reaction speed).
settle_p95_s	95th-percentile settle time — the slow tail of reactions.
unsettled_events	Number of disturbance events that never settled within the 10-minute measurement window.
overshoot_mean_w	Mean overshoot (W): how far grid power swings past zero to the opposite sign after an event.
overshoot_max_w	Worst-case overshoot (W) across all events.
band_crossings_per_h	Sign flips per hour across the ±20 W hysteresis band — oscillation / hunting frequency.
grid_p2p_w	Sustained peak-to-peak grid swing (95th - 5th percentile) over the whole run — oscillation amplitude. Non-zero whenever the loop keeps hunting, including continuous oscillation the step-response metrics (settle/overshoot) miss.
grid_rms_w	RMS grid power (W) over the whole run, transients included — the L2 tracking error: how cleanly the loop held zero, penalising big excursions (overshoot, swings) far harder than a small steady offset. Pairs with battery_travel_w_per_h as the control-effort term.
steady_rms_w	RMS grid power (W) during steady state (excluding the 120 s after each event) — residual jitter when nothing is changing.
mean_abs_grid_w	Mean absolute grid power (W) over the whole run — overall tracking accuracy.
share_imbalance_w	Time-weighted watts misallocated between batteries sharing a phase (sum of each battery's deviation from the even fair share) — 0 when the pool splits load evenly, higher when one battery is left lopsided (issue #523). 0 for scenarios with at most one battery per phase.
avoidable_import_wh	Energy imported from the grid (Wh) the battery could have supplied (it had charge and discharge headroom) — missed self-consumption.
avoidable_export_wh	Energy exported to the grid (Wh) an AC-chargeable battery could have absorbed (it had room and charge headroom) — missed charging.
cost_regret_ct	Money north-star: electricity bill (eurocents, import @ 30 ct/kWh, export @ 8 ct/kWh) over what a perfect-foresight optimal battery would have paid on the same load. Ungameable (both grid directions cost); 0 = matched the optimum. The single number that says how much the controller left on the table.
battery_travel_w_per_h	Total absolute change in battery setpoints per hour (W/h) — control effort / actuator wear; lower is smoother.

Per-scenario tables (31 scenarios)

mixed_cadence/eff — settle 26.1→50.8s, overshoot 232.6→164.5W, RMS 25.6→21.9W

Metric	Base	Head	Δ
settle_mean_s	26.1	50.8	+95%
settle_p95_s	31.6	72.4	+129%
unsettled_events	0.0	0.0	=
overshoot_mean_w	110.0	64.0	-42%
overshoot_max_w	232.6	164.5	-29%
band_crossings_per_h	33.0	24.4	-26%
grid_p2p_w	78.8	160.6	+104%
grid_rms_w	253.5	255.3	+1%
steady_rms_w	25.6	21.9	-14%
mean_abs_grid_w	58.2	64.1	+10%
share_imbalance_w	581.1	223.0	-62%
avoidable_import_wh	34.2	38.5	+13%
avoidable_export_wh	23.9	25.6	+7%
cost_regret_ct	0.83	0.95	+14%
battery_travel_w_per_h	25396.8	31306.0	+23%

mixed_cadence/fair — settle 24.8→47.7s, overshoot 176.3→70.4W, RMS 13.1→12.9W

Metric	Base	Head	Δ
settle_mean_s	24.8	47.7	+92%
settle_p95_s	29.4	67.4	+129%
unsettled_events	0.0	0.0	=
overshoot_mean_w	111.3	30.7	-72%
overshoot_max_w	176.3	70.4	-60%
band_crossings_per_h	27.2	21.6	-21%
grid_p2p_w	27.6	131.9	+378%
grid_rms_w	246.0	246.0	=
steady_rms_w	13.1	12.9	-2%
mean_abs_grid_w	53.5	58.6	+10%
share_imbalance_w	528.9	101.3	-81%
avoidable_import_wh	30.4	33.7	+11%
avoidable_export_wh	23.1	24.9	+8%
cost_regret_ct	0.73	0.81	+11%
battery_travel_w_per_h	23198.8	28457.4	+23%

mixed_cadence_solar/eff — settle 34.6→52.5s, overshoot 351.2→384.7W, RMS 30.3→28.6W

Metric	Base	Head	Δ
settle_mean_s	34.6	52.5	+52%
settle_p95_s	68.8	95.6	+39%
unsettled_events	1.8	2.0	+11%
overshoot_mean_w	104.6	62.4	-40%
overshoot_max_w	351.2	384.7	+10%
band_crossings_per_h	40.0	37.3	-7%
grid_p2p_w	109.1	192.3	+76%
grid_rms_w	267.6	268.4	+0%
steady_rms_w	30.3	28.6	-6%
mean_abs_grid_w	71.0	76.4	+8%
share_imbalance_w	430.0	148.1	-66%
avoidable_import_wh	55.4	60.5	+9%
avoidable_export_wh	51.2	54.1	+6%
cost_regret_ct	1.25	1.38	+10%
battery_travel_w_per_h	30571.0	37320.0	+22%

mixed_cadence_solar/fair — settle 30.8→56.0s, overshoot 169.0→75.4W, RMS 20.3→23.7W

Metric	Base	Head	Δ
settle_mean_s	30.8	56.0	+82%
settle_p95_s	59.0	119.5	+103%
unsettled_events	1.4	1.8	+29%
overshoot_mean_w	86.2	28.7	-67%
overshoot_max_w	169.0	75.4	-55%
band_crossings_per_h	27.2	28.3	+4%
grid_p2p_w	74.5	147.8	+98%
grid_rms_w	257.9	261.6	+1%
steady_rms_w	20.3	23.7	+17%
mean_abs_grid_w	64.9	72.2	+11%
share_imbalance_w	557.9	119.0	-79%
avoidable_import_wh	49.4	56.0	+13%
avoidable_export_wh	47.9	52.4	+9%
cost_regret_ct	1.1	1.26	+15%
battery_travel_w_per_h	24392.2	32435.8	+33%

mixed_venus_b2500/eff — settle 56.2→81.2s, overshoot 220.6→221.9W, RMS 14.9→18.6W

Metric	Base	Head	Δ
settle_mean_s	56.2	81.2	+44%
settle_p95_s	123.7	217.1	+76%
unsettled_events	0.2	2.4	+1100%
overshoot_mean_w	106.1	100.2	-6%
overshoot_max_w	220.6	221.9	+1%
band_crossings_per_h	55.7	49.2	-12%
grid_p2p_w	56.7	67.3	+19%
grid_rms_w	200.6	197.5	-2%
steady_rms_w	14.9	18.6	+25%
mean_abs_grid_w	42.4	43.2	+2%
share_imbalance_w	486.4	263.0	-46%
avoidable_import_wh	41.3	44.2	+7%
avoidable_export_wh	22.3	20.5	-8%
cost_regret_ct	1.06	1.16	+9%
battery_travel_w_per_h	29523.6	39704.6	+34%

mixed_venus_b2500/fair — settle 30.1→75.4s, overshoot 106.9→231.1W, RMS 13.2→22.3W

Metric	Base	Head	Δ
settle_mean_s	30.1	75.4	+150%
settle_p95_s	62.9	191.0	+204%
unsettled_events	0.0	1.4	+1.4
overshoot_mean_w	70.6	110.5	+57%
overshoot_max_w	106.9	231.1	+116%
band_crossings_per_h	29.3	53.3	+82%
grid_p2p_w	42.7	67.1	+57%
grid_rms_w	194.1	189.9	-2%
steady_rms_w	13.2	22.3	+69%
mean_abs_grid_w	37.7	44.5	+18%
share_imbalance_w	621.9	100.7	-84%
avoidable_import_wh	34.4	47.4	+38%
avoidable_export_wh	22.1	19.4	-12%
cost_regret_ct	0.85	1.27	+49%
battery_travel_w_per_h	21395.6	43475.4	+103%

phase_imbalance — settle 44.9→53.4s, overshoot 166.7→145.2W, RMS 30.2→30.3W

Metric	Base	Head	Δ
settle_mean_s	44.9	53.4	+19%
settle_p95_s	106.4	123.6	+16%
unsettled_events	0.6	0.0	-100%
overshoot_mean_w	74.8	64.9	-13%
overshoot_max_w	166.7	145.2	-13%
band_crossings_per_h	79.2	87.6	+11%
grid_p2p_w	42.5	42.4	-0%
grid_rms_w	199.9	199.6	-0%
steady_rms_w	30.2	30.3	+0%
mean_abs_grid_w	40.5	40.4	-0%
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	25.7	25.6	-0%
avoidable_export_wh	14.8	14.8	=
cost_regret_ct	0.65	0.65	=
battery_travel_w_per_h	18583.6	18274.2	-2%

single_venus_d_solar — settle 24.2→24.2s, overshoot 94.4→94.4W, RMS 15.9→15.9W

Metric	Base	Head	Δ
settle_mean_s	24.2	24.2	=
settle_p95_s	27.1	27.1	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	76.8	76.8	=
overshoot_max_w	94.4	94.4	=
band_crossings_per_h	10.7	10.7	=
grid_p2p_w	29.9	29.9	=
grid_rms_w	104.3	104.3	=
steady_rms_w	15.9	15.9	=
mean_abs_grid_w	18.5	18.5	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	17.9	17.9	=
avoidable_export_wh	9.9	9.9	=
cost_regret_ct	0.46	0.46	=
battery_travel_w_per_h	9425.4	9425.4	=

single_venus_d_steps — settle 26.3→26.3s, overshoot 90.3→90.3W, RMS 15.5→15.5W

Metric	Base	Head	Δ
settle_mean_s	26.3	26.3	=
settle_p95_s	33.8	33.8	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	73.3	73.3	=
overshoot_max_w	90.3	90.3	=
band_crossings_per_h	25.8	25.8	=
grid_p2p_w	30.9	30.9	=
grid_rms_w	256.4	256.4	=
steady_rms_w	15.5	15.5	=
mean_abs_grid_w	58.6	58.6	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	32.3	32.3	=
avoidable_export_wh	26.2	26.2	=
cost_regret_ct	0.76	0.76	=
battery_travel_w_per_h	21425.2	21425.2	=

single_venus_d_washer — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 61.0→61.0W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	322.4	322.4	=
grid_p2p_w	192.7	192.7	=
grid_rms_w	61.0	61.0	=
steady_rms_w	61.0	61.0	=
mean_abs_grid_w	42.8	42.8	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	13.3	13.3	=
avoidable_export_wh	8.1	8.1	=
cost_regret_ct	0.33	0.33	=
battery_travel_w_per_h	24204.4	24204.4	=

single_venus_drain — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 907.3→907.3W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	73.2	73.2	=
grid_p2p_w	1598.9	1598.9	=
grid_rms_w	907.3	907.3	=
steady_rms_w	907.3	907.3	=
mean_abs_grid_w	645.3	645.3	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	14.4	14.4	=
avoidable_export_wh	10.2	10.2	=
cost_regret_ct	0.21	0.21	=
battery_travel_w_per_h	4062.0	4062.0	=

single_venus_fill — settle 360.0→360.0s, overshoot 0.0→0.0W, RMS 953.6→953.6W

Metric	Base	Head	Δ
settle_mean_s	360.0	360.0	=
settle_p95_s	600.0	600.0	=
unsettled_events	4.0	4.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	10.1	10.1	=
grid_p2p_w	1713.1	1713.1	=
grid_rms_w	978.6	978.6	=
steady_rms_w	953.6	953.6	=
mean_abs_grid_w	662.8	662.8	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	4.8	4.8	=
avoidable_export_wh	6.7	6.7	=
cost_regret_ct	0.14	0.14	=
battery_travel_w_per_h	3173.0	3173.0	=

single_venus_noisy — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 94.3→94.3W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	1452.6	1452.6	=
grid_p2p_w	296.7	296.7	=
grid_rms_w	94.3	94.3	=
steady_rms_w	94.3	94.3	=
mean_abs_grid_w	79.1	79.1	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	42.0	42.0	=
avoidable_export_wh	37.0	37.0	=
cost_regret_ct	0.96	0.96	=
battery_travel_w_per_h	22758.8	22758.8	=

single_venus_pv — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 60.8→60.8W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	38.5	38.5	=
grid_p2p_w	42.9	42.9	=
grid_rms_w	60.8	60.8	=
steady_rms_w	60.8	60.8	=
mean_abs_grid_w	17.3	17.3	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	9.9	9.9	=
avoidable_export_wh	16.1	16.1	=
cost_regret_ct	0.17	0.17	=
battery_travel_w_per_h	6782.0	6782.0	=

single_venus_solar — settle 26.8→26.8s, overshoot 80.3→80.3W, RMS 17.8→17.8W

Metric	Base	Head	Δ
settle_mean_s	26.8	26.8	=
settle_p95_s	32.6	32.6	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	67.1	67.1	=
overshoot_max_w	80.3	80.3	=
band_crossings_per_h	26.3	26.3	=
grid_p2p_w	41.9	41.9	=
grid_rms_w	108.7	108.7	=
steady_rms_w	17.8	17.8	=
mean_abs_grid_w	21.2	21.2	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	15.5	15.5	=
avoidable_export_wh	16.3	16.3	=
cost_regret_ct	0.34	0.34	=
battery_travel_w_per_h	7331.6	7331.6	=

single_venus_solar_slow — settle 33.9→33.9s, overshoot 68.3→68.3W, RMS 22.8→22.8W

Metric	Base	Head	Δ
settle_mean_s	33.9	33.9	=
settle_p95_s	39.9	39.9	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	32.1	32.1	=
overshoot_max_w	68.3	68.3	=
band_crossings_per_h	5.6	5.6	=
grid_p2p_w	60.6	60.6	=
grid_rms_w	131.7	131.7	=
steady_rms_w	22.8	22.8	=
mean_abs_grid_w	31.6	31.6	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	21.3	21.3	=
avoidable_export_wh	26.1	26.1	=
cost_regret_ct	0.43	0.43	=
battery_travel_w_per_h	6486.0	6486.0	=

single_venus_steps — settle 26.0→26.0s, overshoot 88.0→88.0W, RMS 14.7→14.7W

Metric	Base	Head	Δ
settle_mean_s	26.0	26.0	=
settle_p95_s	32.7	32.7	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	69.7	69.7	=
overshoot_max_w	88.0	88.0	=
band_crossings_per_h	23.8	23.8	=
grid_p2p_w	27.1	27.1	=
grid_rms_w	266.9	266.9	=
steady_rms_w	14.7	14.7	=
mean_abs_grid_w	61.1	61.1	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	34.1	34.1	=
avoidable_export_wh	27.0	27.0	=
cost_regret_ct	0.8	0.8	=
battery_travel_w_per_h	19543.6	19543.6	=

single_venus_steps_slow — settle 40.5→40.5s, overshoot 98.5→98.5W, RMS 14.8→14.8W

Metric	Base	Head	Δ
settle_mean_s	40.5	40.5	=
settle_p95_s	56.2	56.2	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	37.1	37.1	=
overshoot_max_w	98.5	98.5	=
band_crossings_per_h	9.4	9.4	=
grid_p2p_w	78.2	78.2	=
grid_rms_w	331.7	331.7	=
steady_rms_w	14.8	14.8	=
mean_abs_grid_w	88.2	88.2	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	47.5	47.5	=
avoidable_export_wh	40.7	40.7	=
cost_regret_ct	1.1	1.1	=
battery_travel_w_per_h	17699.8	17699.8	=

single_venus_trace — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 278.9→278.9W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	284.8	284.8	=
grid_p2p_w	689.3	689.3	=
grid_rms_w	278.5	278.5	=
steady_rms_w	278.9	278.9	=
mean_abs_grid_w	117.0	117.0	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	81.4	81.4	=
avoidable_export_wh	35.5	35.5	=
cost_regret_ct	1.36	1.36	=
battery_travel_w_per_h	33179.4	33179.4	=

single_venus_washer — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 61.0→61.0W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	320.4	320.4	=
grid_p2p_w	196.1	196.1	=
grid_rms_w	61.1	61.1	=
steady_rms_w	61.0	61.0	=
mean_abs_grid_w	40.6	40.6	=
share_imbalance_w	0.0	0.0	=
avoidable_import_wh	11.3	11.3	=
avoidable_export_wh	9.0	9.0	=
cost_regret_ct	0.27	0.27	=
battery_travel_w_per_h	24290.4	24290.4	=

two_venus/eff — settle 18.9→18.1s, overshoot 150.3→126.1W, RMS 13.9→14.0W

Metric	Base	Head	Δ
settle_mean_s	18.9	18.1	-4%
settle_p95_s	27.1	23.0	-15%
unsettled_events	0.0	0.0	=
overshoot_mean_w	70.3	70.7	+1%
overshoot_max_w	150.3	126.1	-16%
band_crossings_per_h	40.0	33.6	-16%
grid_p2p_w	27.8	26.2	-6%
grid_rms_w	221.6	221.6	=
steady_rms_w	13.9	14.0	+1%
mean_abs_grid_w	43.2	43.1	-0%
share_imbalance_w	272.5	166.4	-39%
avoidable_import_wh	26.0	25.7	-1%
avoidable_export_wh	17.3	17.3	=
cost_regret_ct	0.64	0.64	=
battery_travel_w_per_h	21923.0	22029.2	+0%

two_venus/fair — settle 18.2→18.4s, overshoot 128.7→116.7W, RMS 13.8→13.8W

Metric	Base	Head	Δ
settle_mean_s	18.2	18.4	+1%
settle_p95_s	22.2	24.4	+10%
unsettled_events	0.0	0.0	=
overshoot_mean_w	93.4	94.4	+1%
overshoot_max_w	128.7	116.7	-9%
band_crossings_per_h	20.0	21.6	+8%
grid_p2p_w	23.4	23.1	-1%
grid_rms_w	217.8	217.6	-0%
steady_rms_w	13.8	13.8	=
mean_abs_grid_w	41.5	41.4	-0%
share_imbalance_w	33.7	18.8	-44%
avoidable_import_wh	24.3	24.2	-0%
avoidable_export_wh	17.2	17.1	-1%
cost_regret_ct	0.59	0.59	=
battery_travel_w_per_h	19252.6	19385.4	+1%

two_venus_noisy/eff — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 94.5→94.3W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	2901.2	2898.0	-0%
grid_p2p_w	295.9	294.5	-0%
grid_rms_w	94.5	94.3	-0%
steady_rms_w	94.5	94.3	-0%
mean_abs_grid_w	79.6	79.4	-0%
share_imbalance_w	34.8	26.9	-23%
avoidable_import_wh	43.0	43.6	+1%
avoidable_export_wh	36.6	35.8	-2%
cost_regret_ct	1.0	1.02	+2%
battery_travel_w_per_h	24526.2	23402.6	-5%

two_venus_noisy/fair — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 94.5→94.2W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	2890.4	2893.6	+0%
grid_p2p_w	296.3	294.1	-1%
grid_rms_w	94.5	94.2	-0%
steady_rms_w	94.5	94.2	-0%
mean_abs_grid_w	79.6	79.4	-0%
share_imbalance_w	33.3	26.4	-21%
avoidable_import_wh	43.1	43.6	+1%
avoidable_export_wh	36.5	35.8	-2%
cost_regret_ct	1.0	1.02	+2%
battery_travel_w_per_h	24134.8	22778.8	-6%

two_venus_slow/fair — settle 41.3→41.8s, overshoot 181.1→174.5W, RMS 14.0→14.0W

Metric	Base	Head	Δ
settle_mean_s	41.3	41.8	+1%
settle_p95_s	52.0	52.8	+2%
unsettled_events	0.0	0.0	=
overshoot_mean_w	57.7	56.6	-2%
overshoot_max_w	181.1	174.5	-4%
band_crossings_per_h	11.0	10.4	-5%
grid_p2p_w	74.1	81.8	+10%
grid_rms_w	303.8	303.7	-0%
steady_rms_w	14.0	14.0	=
mean_abs_grid_w	76.8	76.9	+0%
share_imbalance_w	69.9	24.4	-65%
avoidable_import_wh	41.3	41.5	+0%
avoidable_export_wh	35.5	35.3	-1%
cost_regret_ct	0.95	0.96	+1%
battery_travel_w_per_h	18570.4	18563.6	-0%

two_venus_solar/eff — settle 25.5→26.0s, overshoot 417.5→396.6W, RMS 19.6→20.4W

Metric	Base	Head	Δ
settle_mean_s	25.5	26.0	+2%
settle_p95_s	50.2	44.3	-12%
unsettled_events	1.6	1.6	=
overshoot_mean_w	103.6	100.5	-3%
overshoot_max_w	417.5	396.6	-5%
band_crossings_per_h	43.3	40.4	-7%
grid_p2p_w	55.8	60.2	+8%
grid_rms_w	228.7	228.7	=
steady_rms_w	19.6	20.4	+4%
mean_abs_grid_w	51.3	51.9	+1%
share_imbalance_w	160.1	69.3	-57%
avoidable_import_wh	41.9	42.4	+1%
avoidable_export_wh	35.1	35.4	+1%
cost_regret_ct	0.98	0.99	+1%
battery_travel_w_per_h	26691.8	26336.0	-1%

two_venus_solar/fair — settle 22.7→25.9s, overshoot 139.8→151.4W, RMS 19.4→20.4W

Metric	Base	Head	Δ
settle_mean_s	22.7	25.9	+14%
settle_p95_s	37.1	52.1	+40%
unsettled_events	1.2	1.4	+17%
overshoot_mean_w	88.5	98.1	+11%
overshoot_max_w	139.8	151.4	+8%
band_crossings_per_h	30.4	30.7	+1%
grid_p2p_w	54.7	59.8	+9%
grid_rms_w	226.9	225.6	-1%
steady_rms_w	19.4	20.4	+5%
mean_abs_grid_w	50.3	51.1	+2%
share_imbalance_w	219.8	31.3	-86%
avoidable_import_wh	40.4	40.1	-1%
avoidable_export_wh	35.1	36.5	+4%
cost_regret_ct	0.93	0.91	-2%
battery_travel_w_per_h	23851.6	24206.8	+1%

two_venus_trace/eff — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 283.2→283.1W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	306.0	307.6	+1%
grid_p2p_w	745.3	744.9	-0%
grid_rms_w	282.3	282.1	-0%
steady_rms_w	283.2	283.1	-0%
mean_abs_grid_w	122.0	122.0	=
share_imbalance_w	416.1	398.3	-4%
avoidable_import_wh	80.2	80.0	-0%
avoidable_export_wh	41.9	42.0	+0%
cost_regret_ct	2.07	2.06	-0%
battery_travel_w_per_h	48431.2	47999.0	-1%

two_venus_trace/fair — settle 0.0→0.0s, overshoot 0.0→0.0W, RMS 282.6→282.1W

Metric	Base	Head	Δ
settle_mean_s	0.0	0.0	=
settle_p95_s	0.0	0.0	=
unsettled_events	0.0	0.0	=
overshoot_mean_w	0.0	0.0	=
overshoot_max_w	0.0	0.0	=
band_crossings_per_h	300.4	310.0	+3%
grid_p2p_w	725.4	732.2	+1%
grid_rms_w	281.6	281.1	-0%
steady_rms_w	282.6	282.1	-0%
mean_abs_grid_w	121.2	121.1	-0%
share_imbalance_w	63.3	30.3	-52%
avoidable_import_wh	79.7	79.0	-1%
avoidable_export_wh	41.6	42.0	+1%
cost_regret_ct	2.06	2.04	-1%
battery_travel_w_per_h	45944.0	46054.4	+0%

venus_d_plus_c/eff — settle 17.4→20.1s, overshoot 175.4→128.9W, RMS 14.7→14.7W

Metric	Base	Head	Δ
settle_mean_s	17.4	20.1	+16%
settle_p95_s	21.3	31.9	+50%
unsettled_events	0.0	0.0	=
overshoot_mean_w	81.7	78.9	-3%
overshoot_max_w	175.4	128.9	-27%
band_crossings_per_h	33.4	35.2	+5%
grid_p2p_w	28.1	28.0	-0%
grid_rms_w	215.2	214.7	-0%
steady_rms_w	14.7	14.7	=
mean_abs_grid_w	41.9	41.6	-1%
share_imbalance_w	234.0	167.3	-29%
avoidable_import_wh	25.1	24.9	-1%
avoidable_export_wh	16.8	16.7	-1%
cost_regret_ct	0.62	0.61	-2%
battery_travel_w_per_h	22973.2	23208.0	+1%

venus_d_plus_c/fair — settle 18.6→21.6s, overshoot 118.4→121.0W, RMS 14.6→14.6W

Metric	Base	Head	Δ
settle_mean_s	18.6	21.6	+16%
settle_p95_s	24.7	32.5	+32%
unsettled_events	0.0	0.0	=
overshoot_mean_w	84.3	90.6	+7%
overshoot_max_w	118.4	121.0	+2%
band_crossings_per_h	27.0	24.8	-8%
grid_p2p_w	27.0	27.4	+1%
grid_rms_w	211.9	211.5	-0%
steady_rms_w	14.6	14.6	=
mean_abs_grid_w	40.7	40.8	+0%
share_imbalance_w	130.8	30.2	-77%
avoidable_import_wh	24.0	24.1	+0%
avoidable_export_wh	16.7	16.7	=
cost_regret_ct	0.59	0.59	=
battery_travel_w_per_h	20999.4	21212.8	+1%

📊 Open the interactive report — steering-eval-report.html, a single self-contained file (opens in-browser; download it if your browser blocks inline scripts).