With calories and protein matched, the lifters on a sixteen eight window did less total work and their squat went up less

Why they ran it

Almost every time restricted eating trial in trained people runs against a control free to eat differently, so the fasting arm ends up eating less and the result cannot be read. This team matched both arms at a ten percent surplus and 2.2 grams of protein per kilogram per day, supervised the training, and confirmed that calorie, carbohydrate, fat and protein intake did not differ statistically. Whatever separated the groups is then not about intake.

Drawn from the paper's stated aim and its intake reporting.

Seventeen well trained men and women, ten men and seven women, randomised to a sixteen eight eating window or a fed control. Supervised progressive resistance exercise four times a week for eight weeks. Eating windows achieved were 7.9 hours against 13.2 hours.

Calorie, carbohydrate, fat and protein intake did not differ statistically between groups. Both arms were prescribed ten percent above maintenance at 2.2 grams of protein per kilogram per day.

Total exercise volume was significantly lower in the fasting arm, 6,960 repetitions with a standard error of 287 against 7,334 with a standard error of 289.

Subjective daily energy ratings were lower in the fasting arm, by 1.41 at week four with p equals 0.04 and by 1.04 at week eight with p equals 0.06.

Squat one repetition maximum: in the mixed model there was a significant interaction of the fasting group and the post time point at p equals 0.04, with the fasting group improving 4.2 kilograms less, interval minus 8.2 to minus 0.2. The covariate adjusted comparison of final values was minus 4.0 kilograms with a standard error of 1.9, p equals 0.058, which the paper calls a trend.

Leg extension one repetition maximum showed no difference between groups, p equals 0.22.

Fat free mass did not differ between groups. The covariate adjusted comparison of final values was 0.7 kilograms with a standard error of 0.9, p equals 0.46, and both groups gained: 1.8 kilograms at the post time point relative to baseline, interval 0.4 to 3.2.

Fat mass did differ. The fed arm added 1.4 kilograms more, standard error 0.6, p equals 0.04. Mood and sleep ratings did not change in either group.

The numbers

Eating windows achieved	7.9 hours, standard error 0.1, against 13.2 hours, standard error 0.6
Intake	calorie, carbohydrate, fat and protein intake did not differ statistically between groups
Total exercise volume	6,960 repetitions against 7,334, significantly lower in the fasting arm
Subjective daily energy	minus 1.41 at week four, p equals 0.04; minus 1.04 at week eight, p equals 0.06
Squat 1RM, mixed model interaction	fasting group improved 4.2 kg less, 95 percent CI minus 8.2 to minus 0.2, p equals 0.04
Squat 1RM, covariate adjusted final values	minus 4.0 kg, standard error 1.9, p equals 0.058
Fat free mass, covariate adjusted final values	0.7 kg, standard error 0.9, p equals 0.46, no difference
Fat mass	the fed arm added 1.4 kg more, standard error 0.6, p equals 0.04
Leg extension 1RM	no difference, p equals 0.22

Why this might happen

Proposed by the authors This is the explanation the authors offer in their discussion. This study did not test it.

The authors name no physiological mechanism, and the pattern they report points at an experiential one. Subjective daily energy ratings were lower in the fasting arm and total training volume was lower with it. Their conclusion is careful: the approach is viable for building size and strength under a surplus with adequate protein, while the differences in training volume, squat strength, fat mass and energy are, in their words, notable and practically relevant.

Drawn from the paper's results and discussion.

What this does not show

It does not show a compressed window blocks muscle growth. Both groups gained lean tissue and strength. The covariate adjusted comparison of final fat free mass was not different between groups at p equals 0.46, and the discussion says the fasting arm gained an equivalent amount.
Seventeen people split across two arms is small. Every between group estimate here carries a wide interval, and the squat result is the only strength outcome that separated at all.
The squat finding depends on which analysis you read. The mixed model interaction is significant at p equals 0.04 with a 4.2 kilogram gap. The covariate adjusted final value comparison is 4.0 kilograms at p equals 0.058, a trend. Both are in the paper and the honest report includes both.
It does not apply to anyone eating at maintenance or in a deficit. Both arms were ten percent above maintenance with high protein, which is the condition under which a window is least likely to cost anything.

Where this leaves us

Remove the calorie deficit that usually travels with a compressed window, and what is left in trained lifters is less work done and a smaller squat gain.

Seventeen well trained young adults in a supervised eight week block, eating in a surplus with high protein, window starting an hour after training.

The same design at maintenance or in a deficit, where the window has less room to be free.

Caveats worth holding

Seventeen participants across two arms.
Body composition by field methods rather than imaging.
The abstract describes the fat free mass gains as similar and reports a p value of 0.04 alongside them. The full text resolves it: the covariate adjusted between group comparison of final fat free mass is p equals 0.46, and the significant interaction belongs to fat mass. The full text numbers are the ones used here.