Compare two diet periods side by side

New cut starting next week. Before committing, you want to know how the last cut actually went compared to the one before it — adherence, macro averages, weight delta, refeed cadence. Two periods on one card, deltas in plain sight.

What the agent needs to hear

Behind your sentence, the agent works backwards to three pieces: the two periods — named engagements, calendar units, or explicit dates — the dimensions to compare (default: adherence percentage, macro averages, weight delta, refeed compliance), and which baseline reads as « period A » vs « period B », because the direction shapes the sign of every delta.

« Compare last cut to this cut » carries all three by default. « Last cut » resolves to the previous named cut engagement, « this cut » to the most recent one, and the dimensions fall back to the standard set. If you want a narrower read — « just adherence and weight » — say it and the agent trims the card.

Naming periods cleanly

Three ways to name the periods, in descending order of cleanness. By engagement — « the Q1 cut », « last maintenance block », « the cut before this one » — pulls from your engagement history and bounds the dates automatically. By calendar — « April vs May », « Q1 vs Q2 » — the agent resolves against the calendar. By explicit dates — « from Jan 5 to Mar 30 vs Apr 1 to today » — the fallback when the period isn’t named anywhere.

If both periods are the same length, deltas are reported in absolutes. If lengths differ — a 12-week cut against a 4-week maintenance — the agent shows absolute deltas alongside per-week normalizations, so the comparison stays fair even when the windows don’t.

What the comparison actually answers

Comparisons answer questions a single-period read can’t. « Did this cut have better adherence than the last one? » — adherence percentage delta, side by side. « Did I lose more weight per week, or did the rate slow? » — weight delta normalized by duration. « Did I refeed more or less often? » — refeed compliance counts on each side. « What changed in my macros? » — average daily protein, carbs and fat across the two periods.

The shape of your question shapes which dimensions matter. The agent surfaces all four by default; phrase the question narrowly and the card narrows with it.

When the agent gets it wrong

The card is the moment to catch mismatches. Three failure modes show up most. Periods overlap — « no, the Q1 cut ended March 31; Q2 starts April 1, no shared days » — and the agent re-bounds. Wrong dimensions take the foreground — you asked about adherence but the card emphasises macros — narrow it: « just adherence and weight delta ».

The third is direction. If the delta reads the wrong way, name the baseline: « current minus previous, not the other way ». Improvement on a cut is weight loss negative; reading the sign backwards inverts the story.

What makes the comparison worth reading

Three things decide whether this comparison tells you something useful: the periods are comparable in length and intent (a 12-week cut against a 4-week maintenance block reads oddly — pick periods that share a goal shape), the dimensions match the question (asking « did this cut go better » means adherence and weight delta, not absolute calories), and deltas are read in the right direction (improvement on a cut is weight loss negative; improvement on adherence is positive — keep the signs straight).

Read the diff honestly, decide what carried and what slipped, and let the answer set the targets the next plan inherits.