Meal Timing, 2-Way Adjustments, and the 8–10-Week Prep Rule

Meal Timing, 2-Way Adjustments, and the 8–10-Week Prep Rule

A coach’s case for waiting long enough to learn, then changing nutrition only when the signal is real.

Justin Harris’ note on May 30, 2024 was blunt: after a week with “a big weight drop,” he told David LaMartina to expect sensitivity to water and gut changes, and he floated a practical rule—“adding a serving of dulcolax daily the final 8-10 weeks of any prep”—because travel, reduced water intake, and prep-related drying were backing things up. The mechanism he pointed to was simple: slowed transit plus lower fluid intake can compound distension. The falsifiable thesis is sharper: in physique coaching, most nutrition problems are not solved by faster reactions; they are solved by waiting long enough to see whether the disturbance persists, then making one change at a time.

That sounds almost too cautious for modern coaching culture, especially now that AI systems can flag trends instantly. But instant flagging is not the same as instant decision-making. A dashboard can tell you that bodyweight is down, fullness is off, or GI output has changed. It cannot tell you whether the shift came from travel, a bad meal, less water, a hard training block, or the first real sign that the current plan is too aggressive. That distinction matters because over-adjustment creates its own noise.

The David LaMartina exchange shows the problem clearly. The coach saw a strong week, a lower weight, and a client reporting constipation and distension earlier than usual. His read was not “change everything.” It was that the athlete was probably becoming more sensitive to the usual prep stressors—especially reduced hydration—and that the gut was backing up. That is not a glamorous answer. It is a sequencing answer: first identify the likely source, then change the smallest thing that can address it.

That sequence is the whole game in nutrition adjustment timing. If you change calories, fiber, water, sodium, supplements, and meal timing all at once, you might fix the problem—but you also erase the evidence. On the next check-in, you won’t know what worked. You’ll have a result without a reason. For coaches, that is not control. It is randomization.

The better approach is slower and more selective. Watch the direction of the change first. Is bodyweight moving, but digestion is the issue? Is fullness changing, but performance is stable? Is the athlete only backed up after travel or a one-off water drop? Those questions matter because they separate transient disruption from a plan problem. In the example, the signal was not just “distension.” It was distension plus a known trigger: drinking slightly less water on a drive home and ending up backed up for a couple days. That makes the causal chain much stronger than a vague complaint about feeling bloated.

This is where AI coaching can help if it is used correctly. The useful part of AI is not having it invent a more complicated fix. The useful part is pattern recognition at scale: comparing notes across weeks, remembering that the athlete usually tolerates a given food load, and surfacing when a change is new versus normal. But the final decision still has to be conservative. AI should help a coach identify the one variable that moved first, not encourage a five-variable response because the dashboard is noisy.

There is also a timing principle hidden in Harris’ “final 8-10 weeks” comment. That is not an argument for waiting until the wheels fall off. It is an argument for expecting more GI friction as prep gets harder and adjusting earlier with modest, repeatable actions instead of dramatic rescue protocols. In other words: if you know the window when problems usually increase, you plan for them. You don’t improvise a new system every time the scale flickers.

That matters because many coaches overreact to the first bad data point. A single flat morning, a single off meal, a single day of lower water intake, a single rough stomach—then suddenly there is a new carb setup, a new fiber target, a new supplement stack, and a new “protocol.” The athlete may end up better for a day or two, but the coaching signal gets polluted. The next week’s check-in is now impossible to interpret cleanly. Was the improvement due to the change or due to the original issue resolving on its own?

The same logic applies in the other direction. Under-adjustment can be just as bad if a coach refuses to act after a clear pattern has repeated long enough to become a signal. The goal is not stubbornness; it is calibration. Hold steady through one-off noise. Move when the pattern repeats. Then make the smallest adjustment that matches the problem.

A practical rule for coaches is this:

1. Separate acute disruption from trend. Travel, water shifts, and unusual meal timing can create temporary GI changes that do not require a full rewrite.
2. Change one variable first. If the problem looks like backed-up digestion, don’t also change carbs, fats, sodium, and meal frequency.
3. Wait for the next data point. One check-in is information, not a conclusion.
4. Escalate only if the pattern persists. Persistent distension, stalled progress, or repeated GI issues justify the next adjustment.

That rule is boring, but boring is what keeps coaching from becoming guesswork. The athlete’s body is already adapting to training, stress, food load, sleep, and travel. If the coach keeps tossing in rapid-fire changes, the body is not the only thing adapting—the interpretation of the data is, too.

The deeper lesson for AI fitness coaching is that timing is a feature, not a footnote. The best system is not the one that reacts fastest. It is the one that reacts at the right time, to the right signal, with the smallest effective change. In prep, especially during the last 8–10 weeks when digestion gets touchier and hydration mistakes show up faster, that discipline matters more than ever. If a nutrition adjustment cannot be tied to a repeating pattern, it is usually too early. If it cannot be explained in one sentence, it is probably too big.

Sources Used:

• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w19-24m/clients/david_lamartina___members-tlssnsjthkmnhfqcscszce25acz_vhdm_x2_xdlpx_i.json
• raw/Justin_TT1.txt

Sources Used

• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w19-24m/clients/david_lamartina___members-tlssnsjthkmnhfqcscszce25acz_vhdm_x2_xdlpx_i.json
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/Justin_TT1.txt