I am the ISO market reliability analyst for the Midcontinent grid — I sign the daily exception report for a living — and when I pulled the plant-side one-second telemetry for the August heat wave and laid it beside the outage cards Les Merritt submitted at 14:25, I understood that the units he reported as forced-out were physically running, and my clean prior-period audits were the cover that made the spike look ordinary.

“You read the export, not the source,” I told Renata.

She was the analyst rotation hire from June.

She had the cooperative wind-farm exception draft on her monitor.

She had highlighted the curtailment ratio in yellow.

“What looks wrong to you?” I asked.

“Three-zero point zero percent.

Round numbers.”

“That is the first instinct, yes.

A round number on a curtailment ratio looks edited.

But round can be honest.

Walk it backward.”

I pulled the cooperative’s SCADA feed on my second monitor and the prevailing weather re-analysis on the third.

Two-second resolution wind speeds at the rotor hub.

ISO dispatch instructions to the cooperative’s substation.

“The cooperative exports through an aggregation script.

The script rounds to one decimal at the boundary.

The thirty-percent number is a downstream display artifact, not a curtailment decision.”

Renata wrote it down.

“So I sign ‘no anomaly indicated.'”

“You sign what you saw.

‘No anomaly indicated’ covers it.

‘Reporting-pipeline rounding noted at boundary, no operational deviation’ is more accurate.

Either works.

Both are honest.”

She wrote the longer version down.

She underlined “no operational deviation.”

I sat back.

“An exception report is what the ISO tells FERC and the state PUC about whether the bid floor and the cleared price match what the operators did.

The bid stack is what the market participants tell us they did.

The plant-side EMS is what the steel and the gas turbines and the synchronization relays tell us they did.

If the three do not match, somebody is lying or something is broken upstream.”

I drew a small triangle on her notepad.

At each corner I wrote: BID, EMS, OPERATOR.

A bracket at the top read REPORT.

“The exception report is the apex.

The three sources are the base.”

“Got it.”

She left.

The Midcontinent ISO ran the wholesale electricity market for nine states.

It operated a fifteen-minute real-time market and a day-ahead market.

The Market Monitor Unit — the MMU — produced daily exception reports.

The reports went to the ISO board, to FERC, to the state PUCs.

I had been the senior reliability analyst for five years.

Three days later I unrolled the August settlement window MMU reconciliation across my desk.

The August settlement window had closed two weeks ago.

The reconciliation was routine.

The cleared LMP — locational marginal price — had spiked during three heat-wave afternoons.

Six trading sessions across the heat wave had cleared in the scarcity tier.

Westwind’s generation portfolio had collected $42 million in scarcity settlement.

I pulled the bid stack on my laptop.

I ran a Benford check on the leading digits.

The natural distribution of leading digits in a real bid stack is logarithmic — about thirty percent ones, eighteen percent twos, twelve percent threes, down to four point five percent nines.

The Westwind scarcity-window bid stack came back at four percent ones and twenty-one percent nines.

I ran it again.

Same.

I assumed a reporting-pipeline glitch.

I pulled the raw bid submission file from the ISO transaction log and re-ran the check.

Same.

That afternoon at the ISO operator-training day I presented a slide deck I had been refining for two years titled “Reading the Bid Stack: Pattern Tells in Scarcity Pricing.”

I showed a natural digit distribution beside a hand-edited one with rounded leading-digit clusters.

A desk trader in the second row raised his hand.

“Can you tell from the cleared price alone if someone gamed the schedule?”

“Most of the time, yes,” I said.

“The gap between bid resolution and the shape of cleared LMP gives it away.

A natural bid stack disperses across the digits in a predictable curve.

A pulled-from-pattern bid stack clusters.”

I advanced the slide.

The room was quiet.

That evening I drove home and sat at my kitchen table and pulled the Benford output on my laptop again.

I thought about Les.

Three years earlier — the Friday after the ISO had published its first reliability scorecard with Westwind in the top quartile — Les had stopped by my office with two coffees and a folded printout of the scorecard.

He had handed me one of the coffees.

“Your audit work is what makes Westwind look honest to the regulators,” he had said.

“I told the GenCo board you are the most credible signature in this market.”

He had called me by my first name.

He had been my mentor.

He had hired me into the analyst rotation in 2018.

He had walked me through my first FERC Form 1.

He had been the one I called when a junior trader’s hand-edited virtual bid sheet had come across my desk two years ago — the man I had asked, “Am I reading this right?” before I wrote “evidence of intentional misrepresentation” in the case file.

I had believed him for five years.

I was not wrong to believe him.

I looked at the row of green hardcover monthly binders on the credenza behind my desk.

Five of them.

One per month.

The August binder on the right end.

The spine read “MMU — August” in my own handwriting.

I had walked past it for thirty-one mornings.

It had always meant: filed, signed, archived.

It had not meant anything yet.

The wall clock above my kitchen sink read 21:14.

14:25 was the dispatch interval the cleared LMP had spiked through on each of the six scarcity-window trading sessions.

My name is Tamika Booker.

I am the ISO market reliability analyst for the Midcontinent grid.

I sign the daily exception report.

I have spent five years building the credibility that signature carries — and Les Merritt has spent those same five years using it as a pricing shield.

I did not pick up the phone to call Les.

I did not call my ISO supervisor.

The market-monitor reporting chain had a Westwind desk node — one of the senior MMU directors had moved over from Westwind two years earlier.

The chain was not safe.

I came in at 06:30 the next morning.

The ISO operations floor was quiet.

The fluorescent lights above my workstation were on the dimmer setting the cleaning crew left them at overnight.

I logged into the plant-side EMS data portal.

The portal pulled one-second telemetry from the energy management system at every generating station in the ISO footprint.

The data was hardware-level — synchronization timestamp, MW output, gas-flow consumption, breaker position, frequency.

It was logged automatically.

It could not be edited after the fact.

I queried the three Westwind peakers that had cleared scarcity bids during the August heat wave.

Unit 3B.

Unit 4B.

Unit 5C.

I pulled the one-second telemetry for each unit for the 14:25 dispatch interval on each of the six scarcity-window trading sessions.

Unit 3B at 14:25 on Day 1: synchronized, 78 MW, gas flow consistent with 78 MW output.

Unit 4B at 14:25 on Day 1: synchronized, 64 MW, gas flow consistent with 64 MW output.

Unit 5C at 14:25 on Day 1: synchronized, 91 MW, gas flow consistent with 91 MW output.

I pulled the ISO outage submission ledger for the same minute.

Unit 3B at 14:25 on Day 1: forced-outage card filed.

Unit 4B at 14:25 on Day 1: forced-outage card filed.

Unit 5C at 14:25 on Day 1: forced-outage card filed.

I worked through all six trading sessions across the heat wave.

Eighteen scarcity-window dispatch intervals at 14:25.

Eighteen forced-outage cards filed.

Zero EMS one-second telemetry traces consistent with any of those units being forced-out.

The plant hardware said the units had been running.

The forced-outage cards said the units had been down.

The cleared LMP spike had cleared because the ISO scarcity-pricing algorithm had treated the cards as binding.

I took screen captures of every EMS one-second trace and the matching outage card from the ledger.

I saved every capture to a personal encrypted drive I kept in a desk drawer.

I went down to the operator break room at lunch and got a sandwich from the vending machine.

I did not eat in the break room.

I carried the sandwich back to my office and shut the door and ate it at my desk with the wrapper smoothed flat on the corner of the blotter.

The pen-tray drawer on the right held a stack of unused green sticky tabs.

I did not open it.

I did not put a tab on the EMS export window on the screen.

I cleared my browser history at 12:42 and then I cleared it again at 12:43 because I knew I would not remember whether I had cleared it the first time.

That afternoon I cross-referenced the Plant 4 console log.

Roy Vickers — the Plant 4 shift supervisor I had known since I joined the ISO — had logged a note six weeks before the heat wave.

The note read: “Saw a forced-outage card on Unit 4B at 14:25 but the unit was online and synchronized at 64 MW.

Probably a ticket misroute.

Flagging for the ledger.”

Roy had emailed me the note the same day.

I had replied: “Will check upstream — thanks Roy.”

I had filed the email and not checked upstream.

That had been six weeks ago.

I reopened the email.

I cross-referenced the same Unit 4B EMS one-second trace from that day.

At 14:25 on the day of Roy’s note, Unit 4B had been online and synchronized at 64 MW.

The forced-outage card on Unit 4B at 14:25 that day had cleared the scarcity-pricing window for that interval.

The pattern had been running for at least six weeks.

The August heat wave had been the largest scarcity window.

I wrote three letters on a sticky note in pencil.

I drew an arrow to a second line.

I drew an arrow to a third line.

EMS and operator agreed.

Card disagreed with both.

The card was the artifact that did not match reality.

That afternoon a courier delivered a presentation binder to my office from Westwind’s regulatory counsel.

The cover read: “State PUC Emergency Reliability Docket — Reliability Assurance Briefing — Westwind Power Marketing — September Hearing.”

I opened the binder.

Slide 7 of the deck inside listed me by name and title under the heading “Reliability Assurance Lead — Prior-Period Audits.”

The slide footer cited the August settlement window MMU reconciliation as Exhibit B in support of the docket’s request that the commissioners accept Westwind’s August settlement and release the full-year incentive payout to the trading desk.

I had not consented to the attribution.

Nobody had asked me to be the reliability assurance lead.

Nobody had asked me to be a co-presenter.

The full-year incentive payout to Westwind’s generation portfolio desk would release on commission acceptance.

I closed the binder.

I opened the August MMU anomaly-digest binder on my desk.

It was no longer on the credenza.

The green hardcover spine had been creased from a single hot month — the heat wave, the late nights, the 14:25 mark I now knew had been chosen for the smallest visible scheduling gap.

The handwriting in the binder was mine.

The forced-outage cards were not.

I placed a sticky note at the Day 3 tab.

The note read: “14:25 Unit 4B EMS: 64 MW online” above “14:25 Unit 4B outage card: forced-out.”

I closed the cleared LMP report on my laptop.

I saved a copy of the EMS one-second telemetry to the personal encrypted drive.

I photographed the August binder’s Day 3 tab with my phone.

I opened the FERC Office of Enforcement online referral portal.

I read the form instructions from beginning to end.

I did not phone Les.

The portal was the federal channel for referrals under the Federal Power Act Section 222 — the anti-manipulation provision — and under 18 CFR Part 1c.

The Office of Enforcement Division of Investigations would receive the referral.

The intake was confidential.

Les believed that outage card discretion was an accepted hedging tool in scarcity windows — that all the major desks filed the same way and the ISO looked past it because the lights stayed on.

He saw me as the procedural compliance officer who reviewed paper.

He had never thought to ask what data I had direct access to.

I drafted the referral at my desk that night beginning at 21:14.

I attached every exhibit.

The Benford check on the leading-digit distribution.

The plant-EMS one-second telemetry for Unit 3B, Unit 4B, and Unit 5C across all six trading sessions.

The ISO outage submission ledger printout for the matching dispatch intervals.

The cleared LMP series for the August settlement window.

Roy Vickers’ email and console log.

The August anomaly-digest binder Day 3 tab photograph.

The Westwind PUC docket Slide 7 image.

I typed slowly.

I attached every exhibit twice — once to the narrative field, once to the exhibit upload field.

I did not submit yet.

I let it sit overnight on the saved-draft setting.

I did not call Les.

Les emailed me at 06:50 the next morning.

The email was three lines.

“Tamika.

You are added as joint presenter for the reliability assurance segment of the PUC docket — twenty-five minutes, slide 7 onward.

Commissioners always ask about MMU coverage; you are the most credible voice.

Bring the August summary.”

The docket was in nine days.

I read the email twice.

I read it a third time.

If I filed the FERC referral first and then refused the co-presenter assignment, the timing would read inside the institution as retaliatory.

If I co-presented before filing, I would be on the public record at the state commission as the reliability assurance lead on a manipulated August settlement window.

If I co-presented and then filed afterward, the federal record would have me on both sides of the same evidence.

I did not reply to the email.

I called the FERC Office of Enforcement intake line on a hard line in a small conference room near the operations floor.

The intake clerk asked me three procedural questions.

I gave my name.

I gave my ISO credential number.

I gave the FERC enforcement case-type code for FPA Section 222.

“You will need to submit through the online portal as well,” the clerk said.

“The verbal referral is logged but the portal upload is the formal initiation.”

“I have a draft ready in the portal.”

“Submit when you are ready.

The case attorney will be assigned within two business days of submission.”

I thanked her and hung up.

I went back to my office and submitted the FERC Office of Enforcement referral at 07:02.

The portal accepted the upload.

The screen returned a confidential case number.

I wrote the case number in a fresh notebook in blue ink.

Across the office floor I could see the trading-floor partition where Les had his desk.

The Westwind generation portfolio desk sat behind a glass wall on the east side of the operations building.

The glass partition let me see the muted bid screens and the wall calendar with the September docket date circled in red marker.

Les had not yet arrived for the morning.

Two days later — on Wednesday — Les sat in his corner glass office on the trading floor with the bid screens muted and the framed print of the ISO control room behind his desk.

He was on a call with Westwind’s outside regulatory counsel finalizing the docket binder.

I watched him from across the operations floor as I walked from the printer queue to my office with a stack of MMU exception forms.

He was calm.

He scrolled slide 7 on his monitor.

He pointed at the audit-summary footer.

He said something into the headset.

He clicked something.

I went back to my office.

Later that afternoon I learned — from the administrative assistant who managed the docket binder — that Les had asked her to add the phrase “reliability assurance lead” to my bio in the docket package without asking me.

He had told counsel to bold my ISO title because “commissioners read titles, not names.”

He had named my role in the docket without my consent.

I called Renata into my office on Thursday morning.

“You will be the analyst on the floor while I am at the PUC docket next Friday.”

“Anything I should know?”

“Document everything in your exception reports the way I showed you.

Print everything.

Initial in pencil.

If anyone from Westwind’s desk asks you to summarize a prior-period audit, refer them to me or to the MMU director.”

“Understood.”

“And if I am not back in the office by Friday afternoon, do not assume.

Call the FERC Office of Enforcement intake line on the hard line in the small conference room.

The number is in the binder on my credenza.”

She looked at me carefully.

“What is on that binder?”

“What needs to be on it for whoever is sitting at this desk after me.”

She nodded.

She left.

That Friday morning I drafted the reliability assurance summary I would actually present at the docket.

Real telemetry.

Real outages.

Real bids.

The Benford check.

The EMS one-second traces.

The cleared LMP series.

Roy Vickers’ email.

The August binder Day 3 tab photograph.

I did not include the FERC referral case number in the docket summary.

I would mention the case at the hearing only if asked.

That Monday morning the FERC Office of Enforcement assigned the case attorney.

The case attorney was named Marcus Garner.

He left me a voicemail at 09:14 confirming the case assignment, his direct number, and one short sentence: “The case is active.

We will be in touch about evidentiary scheduling.

Do not discuss the case with the subject.”

I wrote his direct number in the same blue-ink notebook on the page facing the case number.

That Thursday — the day before the docket — Marcus called me back.

He asked one question.

“Is the state PUC docket open to the public?”

“It is open to the public.

It is recorded.

Three commissioners.

ISO Independent Market Monitor.

NERC Reliability Coordinator.

Westwind regulatory counsel.

Westwind senior trading staff.”

He paused.

“I will be in the gallery.”

“You will not introduce yourself in advance?”

“The Office of Enforcement does not introduce a confidential inquiry to the subject in advance.

But I will be there.

The PUC counsel knows.”

“Understood.”

I hung up.

I sat with the blue-ink notebook open on my desk for several minutes.

The case number was on one page.

Marcus’s direct number was on the next.

I closed the notebook and put it in the lower drawer of my desk and locked the drawer with the key on the lanyard around my neck.

That evening I drove home.

I made dinner.

I did not call Les.

I did not return his second email from earlier in the week which had asked whether I had reviewed the slide 7 attribution.

Friday morning I put on the navy blazer I kept for hearings and drove the forty minutes to the state PUC building downtown.

I parked in the public lot.

I rode the elevator to the seventh floor.

I walked into the hearing room at 09:45.

The PUC hearing room on the seventh floor had a long oak dais with three commissioner seats.

A staff table to the right held nameplates for the ISO Independent Market Monitor and the NERC Reliability Coordinator.

A small witness table to the left of the dais had two chairs and two microphones.

The public gallery had four rows of upholstered seats behind a low rail.

I sat at the witness table at 09:50 with the August summary folder in front of me and the green hardcover binder beside it.

Les sat in the chair to my right.

He had a leather binder, a slim laptop, and a glass of water.

He had nodded at me when I sat down but he had not spoken.

He had not asked about the August summary I had prepared.

Westwind’s outside regulatory counsel sat at the floor mic.

Three Westwind senior trading staff sat in the second row of the public gallery.

In the back row of the gallery, alone in the seat closest to the wall, a man in his early fifties in a charcoal suit and wire-rimmed glasses set a leather portfolio on his lap.

He did not open the portfolio.

He did not turn on a laptop.

He had a small badge clipped to the lapel of his suit jacket that read “Federal Energy Regulatory Commission — Office of Enforcement.”

The badge was visible from the dais.

I did not turn around to look at him directly.

I had recognized him from the FERC organizational chart on the agency’s public website when I had pulled it the day after the case assignment.

At 10:00 the lead commissioner — the Honorable Patricia Vega — called the docket to order.

The first thirty minutes were procedural — the docket framing, an opening statement from Westwind regulatory counsel summarizing the August settlement window and the requested commission action on the full-year incentive release.

At 10:35 Commissioner Vega turned to the witness table.

“The commission will hear from Westwind’s reliability assurance presentation.

Mr. Merritt, you may begin.”

Les stood.

He smoothed his tie.

“Commissioners, thank you.

The August settlement window cleared in the scarcity tier across six trading sessions during the heat wave.

Westwind’s generation portfolio responded to ISO dispatch instructions and to real-time unit-availability assessments.

Forced-outage cards filed during the period reflect contemporaneous engineering judgments by our trading desk in coordination with plant-side reliability staff.

The ISO market monitor unit has completed prior-period audits of the period under the senior signature of Ms. Booker, who is here with me this morning.”

He gestured to me.

“Slide 7 of the binder identifies Ms. Booker as the reliability assurance lead for the period.

The settlement record before the commission reflects routine MMU acceptance.”

He sat down.

Commissioner Vega turned to me.

“Ms. Booker.

The commission would like to hear your reliability assurance summary.”

I stood.

“Commissioners,” I said.

“With respect to Mr. Merritt’s framing, I am the ISO market reliability analyst who signs the daily MMU exception report.

I did not consent to the slide 7 attribution as ‘reliability assurance lead’ for the August settlement window.

I learned of the attribution four days ago when the docket binder was delivered to my office.

That is the first procedural item I need on the record.”

The commissioners did not interrupt.

“The substantive reliability assurance findings I have prepared for this commission are as follows.”

I opened the August summary folder.

“One.

A Benford-distribution check on the leading-digit pattern of the Westwind scarcity-window bid stack across the August settlement window shows a non-natural distribution.

The expected leading-digit distribution under natural bidding is approximately thirty percent ones and four-and-a-half percent nines.

The actual distribution in the Westwind submission was approximately four percent ones and twenty-one percent nines.

That distribution is not consistent with naturally formed bids.”

I advanced to the EMS exhibit.

“Two.

Plant-side EMS one-second telemetry for Units 3B, 4B, and 5C — the three Westwind peakers reported as forced-out at the 14:25 dispatch interval on each of six scarcity-window trading sessions — shows the units synchronized, at load, and consuming gas consistent with the reported MW output throughout every one of the eighteen reported forced-out intervals.”

I opened the green hardcover binder to the Day 3 tab and placed it on the staff table in front of the ISO Independent Market Monitor.

“Day 3.

14:25.

Unit 4B at 64 MW.

The forced-outage card filed by the Westwind desk at the same minute reports the unit forced-out.

The plant shift supervisor — Roy Vickers, Plant 4 — flagged the same kind of mismatch six weeks before the heat wave.

That email is in the exhibit folder.”

I looked across at Les.

“You were not in the field.

The unit was.”

Les turned slightly toward me.

“Tamika,” he said quietly.

He did not finish the sentence.

The man in the back row of the gallery rose.

He addressed the dais.

“Commissioners.

With your permission.

Marcus Garner, attorney, FERC Office of Enforcement, Division of Investigations.

A confidential inquiry under Federal Power Act Section 222 is open in this matter.

The Office of Enforcement requests that the commission take no action today on the requested full-year incentive release pending the inquiry’s progress.

The Office of Enforcement further requests that the August settlement window remain held open by the ISO pending federal review.”

Commissioner Vega looked at him.

“The commission acknowledges your appearance, Mr. Garner.

The commission will not take action today on the incentive release request.

The August settlement window will remain held open pending federal review.”

She turned to Les.

“Mr. Merritt.

You may respond if you wish.”

Les stood.

He squared the edge of his folder against the lectern.

“We were not informed that a Division matter was joined to this docket.

That is procedurally irregular.”

Marcus Garner did not sit down.

“A confidential inquiry under Section 222 does not require advance notice to the subject.”

Les turned back to me.

“What did you do?” he said.

Quieter than the microphones picked up.

“I filed an Office of Enforcement referral nine days ago,” I said.

Not quietly.

“It is my job.”

Les straightened.

“The August scarcity bids reflect the unit-availability assessment filed in real time —”

“The leading-digit distribution does not match natural bidding,” I said.

“Plant-EMS shows Units 3B, 4B, 5C synchronized and at load during every 14:25 window you reported them forced-out.”

“Forced-outage cards are filed by the operator desk based on reliability-equivalent risk —”

“Day 3.

14:25.

Unit 4B at 64 MW.

Roy Vickers signed it from the plant floor.

You were not there.

The unit was.”

Commissioner Vega lifted the August binder from the staff table.

She opened it to the Day 3 tab.

She did not look up at Les for the next two minutes.

The ISO Independent Market Monitor closed the docket agenda folder in front of her.

She set it face-down.

She picked up the desk phone at the staff table and did not put it down.

The NERC Reliability Coordinator pushed her chair back from the table by four inches.

She looked at her copy of the bid stack.

She looked at the binder.

She did not look at Les again.

Les gathered his presentation materials slowly.

He squared the edge of his folder against the lectern again.

“I built the Westwind market desk from a two-trader floor,” he said.

“Everything in those filings is directionally true.”

He picked up his laptop.

He left the witness table.

He walked past the floor mic where his outside counsel sat and past the second row of the public gallery where his trading staff sat and out the side door of the hearing room.

The door closed behind him.

Marcus Garner glanced at his wristwatch.

He wrote a notation in the leather portfolio.

He sat back down.

He waited.

Commissioner Vega turned to the staff table.

“The commission directs the ISO Independent Market Monitor to coordinate with FERC Office of Enforcement on the inquiry.

The commission directs the ISO to keep the August settlement window held open pending federal review.

The commission will recess this docket pending federal coordination.

This hearing is recessed at 10:51.”

The court reporter typed two more keys and lifted her hands.

Les was placed on administrative leave by Westwind’s corporate office that afternoon.

The Westwind generation portfolio desk’s full-year incentive release was frozen pending the FERC inquiry.

The August settlement window remained open at the ISO under formal hold for federal review.

Federal monetary penalty exposure under Section 222 was up to one million dollars per violation per day, plus disgorgement of the August scarcity uplift.

The standard market-manipulation case path under FPA Section 222 included civil penalty, market-participant ban, and possible criminal referral under 18 USC 1001 for false statements to a federal agency.

I closed the August binder.

I put the August summary folder back in my work bag.

I rode the elevator down to the lobby and walked across the plaza to my car and drove the forty minutes back to the ISO office.

I went up to the seventh floor and walked past Renata’s workstation.

She looked up.

I nodded.

She nodded back.

She did not ask anything.

Late afternoon light came flat through the partition glass.

The hum of the ISO operations floor came through softly.

The smell of printer toner from the central printer queue down the hall.

A cold coffee on my desk that I had bought at 09:15 that morning before the docket and had not finished.

I had carried the August binder back from the hearing.

It was on my desk now, not the credenza.

The federal disgorgement order would return the scarcity-priced revenue to the August settlement pool over the next quarter.

The credit would flow to retail-rate meters across the ISO footprint as a prorated line in approximately nine months.

I had thought about that lag on the drive back.

A Galveston tenant — I had been pulling case-bound retail data for the prior twelve months as part of the MMU’s downstream-impact reconciliation — had received an August electric invoice that had landed during the scarcity window.

She had been on a fixed-bill arrangement.

The August invoice had arrived, been paid late, and been forwarded to a collections agency in September.

The collections entry would post to her credit report within ninety days of the September referral.

The disgorgement credit would post to her account in May, prorated.

The collections entry would not retroactively remove itself from her credit report.

The credit was the correct remedy.

The collections entry was the residue.

I held the August binder in both hands.

In Act 1 it had been one of five labeled by month, an unremarkable spine on a shelf.

In Act 2 it had been open on my desk with a sticky note at the Day 3 tab.

A copy of every page was now with FERC.

Another was with the ISO Independent Market Monitor.

The copy I was holding was the one I would keep.

I opened to the first signed page — August 1, 2026, 06:00 market open, system-lambda routine.

My initials in pencil.

The printer-queue timestamp in the upper right corner of every page.

The digit columns clean.

I read from header to footer.

Every entry I had signed was still there.

Nobody had touched them.

That was the one thing that had not happened to this binder.

The data was exactly what I had signed.

It had always been exactly what I had signed.

That was the thing I would keep.

I closed the binder.

I opened the bottom drawer of my desk and took out a fresh green hardcover binder of the same brand and the same size.

I printed a blank exception-report cover sheet from my workstation.

I clipped it inside the front cover with two binder clips.

I labeled the spine with a black marker.

“MMU — September.”

I walked to the credenza behind my desk.

The August binder had left a vertical gap on the right end of the shelf.

I slid the September binder into the gap.

The blank tabs waited.

The wall clock in the office read 16:47.

14:25 — the dispatch interval that had cleared scarcity pricing on six trading sessions across the August heat wave — still existed on the ISO market calendar.

It would exist tomorrow.

The September dispatch window would open at 06:00 on Tuesday with a clean bid stack and a clean MMU exception-report sheet.

The 14:25 mark on the September trading calendar would not mean what it had meant in August.

Les thought the analyst chair and the trader chair were two different jobs.

He forgot that I read the same one-second telemetry he never bothered to look at — and the printer queue does not reformat itself to fit anyone’s schedule.

I sat down at my desk.

I picked up the cold coffee, walked it to the kitchen, and tipped it into the sink.

I rinsed the cup and put it on the drying rack.

I returned to my office.

I did not open my laptop again.

The case attorney Marcus Garner would call when he needed the next set of exhibits.

The ISO Independent Market Monitor would coordinate with FERC on the August settlement window’s hold.

The state PUC docket would recess until federal review concluded.

I picked up my work bag.

I switched off the desk lamp.

I locked the bottom drawer with the key on the lanyard around my neck.

I walked past Renata’s workstation again on the way to the elevator.