The Night I Kept a Losing System Too Long

It was a Tuesday in late October, somewhere around two in the morning, and I was standing at the edge of the main floor of my Las Vegas flagship watching a shift change that should have taken eleven minutes run past thirty. I had a bourbon in my hand that had gone warm. I had not taken a sip. I was watching the same handoff problem I had watched for the better part of eighteen months, and I was still telling myself it was a calibration issue — that the system would settle, that the people would adapt, that patience was the correct instrument here.

It was not patience. I know that now. What I was doing that night, and for the months before it, was something quieter and more stubborn: I was protecting a decision I had already made. The system was mine. I had designed the core logic of it myself, back when the floor was smaller and the logic actually held. Letting go of it meant admitting that what had worked at one scale had quietly stopped working at another — and that I had watched it fail, repeatedly, without acting. That is a harder thing to admit than most people suppose, even in private.

I stayed on that floor until nearly four. I did not fix anything that night. But I did finally stop pretending I was watching a calibration problem.

What I Was Actually Seeing on That Floor at 2 a.m.

The system in question was an operational handoff protocol — the structured way information moved between outgoing and incoming shift supervisors across the floor. When I had built it, the floor had three pit sections and a staff small enough that informal trust filled the gaps the protocol left open. People knew each other. They filled in what the paperwork missed because they had been standing next to each other for years.

By the time I was watching it fail on that October Tuesday, the floor had expanded twice. New sections, new staff, new pit bosses who had not grown up inside the culture of the original team. The informal trust was gone. The protocol, unchanged, was now a skeleton without muscle — the form of a handoff without the substance of one. Information was dropping. Decisions were being made twice, or not at all. And the drag on the floor was real: slower responses, small service failures accumulating, supervisors visibly frustrated with a process that felt like it was working against them.

"Mr. D, I've been doing this handoff the same way for eight months and I still don't know if the next guy actually reads what I leave him." — Marco, pit supervisor, third section

Marco said that to me one evening about six weeks before the Tuesday I am describing. He said it plainly, without complaint in his voice — just a man reporting what he observed. I told him I would look into it. I had already been looking into it for a year. What I had not done was change anything.

The Reasoning I Used to Stay Still

I have thought about this enough to be honest about the structure of my own resistance. It was not laziness — I was actively managing the problem in my head, running through variations, drafting revisions I never implemented. It was not ignorance — Marco was not the first person to tell me what he told me, only the most direct. What it was, I think, was a particular kind of attachment that masquerades as strategic patience.

The reasoning went roughly like this: the system had worked before, therefore the problem was environmental, therefore the solution was to adjust the environment rather than the system. Give the new supervisors more time to absorb the culture. Add a layer of verbal confirmation on top of the written protocol. Wait for the floor to settle into itself. Each of those adjustments bought another few months of not having to say the original design was inadequate.

There is a version of patience that is genuinely strategic — the kind I have written about in the context of holding a position when the evidence still supports it, even when the room disagrees. That patience has served me well and cost me less than impatience ever did. But this was not that. This was patience used as a delay mechanism. The evidence had shifted. I had not shifted with it.

The tell, when I finally let myself read it, was in the supervisors' faces during the handoff I watched that Tuesday. They were not confused. They were resigned. Confusion means the system is new. Resignation means the system has been broken long enough that people have stopped expecting it to work. That is a different problem entirely, and it is harder to fix because you are not just fixing the protocol — you are also repairing the trust that the protocol will ever be worth following.

What Those Eighteen Months Actually Cost

I want to be precise here, because precision is the only honest way to account for a mistake.

The operational drag was real and measurable — slower floor responses, duplicated decisions, supervisors spending mental energy navigating a broken process instead of reading the room. That is not nothing on a floor that runs every hour of every day. But the operational drag was not the deepest cost.

The deeper cost was Marco, and the three or four supervisors like him — experienced, observant people who had flagged the problem clearly and watched nothing happen. When you do not act on what good people tell you, you are not just failing to fix the problem they named. You are teaching them something about whether their observations matter. I had spent years building a floor culture where the people closest to the action felt authorized to speak up, because what gets seen at floor level rarely makes it up the chain intact, and the gap between those two views is where most operational failures live. I had worked to close that gap. And then I had spent eighteen months demonstrating, through inaction, that closing it was optional.

Marco did not leave. He is still there, still one of the sharpest supervisors on the floor. But I noticed, in the months after that Tuesday, that his reports became slightly more formal — a little less of the plain, direct observation he had offered me that evening. I cannot prove causation. I suspect it anyway.

I also lost something in myself that is harder to name — a clean relationship with my own judgment. There is a rule I hold about acting when the evidence is clear, and I had bent it slowly, over eighteen months, until it was nearly unrecognizable. Restoring that kind of internal clarity takes longer than fixing a handoff protocol. It requires doing the right thing enough times, in enough subsequent situations, to trust yourself again. That work is unglamorous and it does not show anywhere on a balance sheet.

The Thing I Still Believe, Held a Little More Carefully Now

I still believe in systems. I believe in structure over improvisation, in designed processes over informal ones, in the unglamorous machinery of structural advantage — the kind that works not because of brilliance in the moment but because of decisions made carefully in advance. None of that changed.

What changed is my understanding of the maintenance that systems require — not just technical maintenance, but the discipline of watching them honestly. A system that worked at one scale is not the same system at a larger one. The logic may be identical; the conditions are not. And the people operating inside the system will tell you, usually clearly, when the fit has broken. The question is whether you are listening or whether you are, as I was, listening selectively — hearing the reports but filtering them through a prior conclusion you have already committed to.

Jeff asked me about this, later, when I told him the story. He had a way of reducing things to their plainest form that I had noticed from the first night I met him. He said: "So you knew, you just didn't want to know you knew." I told him that was about right. He nodded and did not press further. That is one of the things I respect about him — he does not need to be right out loud.

What I hold now is something like this: patience and avoidance feel identical from the inside, especially when the thing you are avoiding is your own prior judgment. The difference only becomes visible in retrospect, and usually only when someone like Marco has already been waiting long enough to go quiet.

I replaced the protocol in November. The new version took three weeks to design and two weeks to implement. The handoffs now run under twelve minutes on every shift. I do not know whether I would have designed a better system if I had acted sooner, or whether I needed eighteen months of watching it fail to understand exactly what was wrong. I suspect the honest answer is that I needed about four months, and spent the other fourteen protecting something that no longer deserved protecting. Whether that ratio is typical, or whether I am unusually slow to let go of my own designs — that is a question I find myself returning to more often than I expected.

Note: Victor Draemont is a character, and these notes are written in his voice. Nothing here is financial, investment, tax, or legal advice, and nothing here is a recommendation to buy, sell, or gamble on anything.