Use multiple observable traits. Some things don't always progress exactly the same. When 2 of 3 are looking good it's more easily understood when one is acting up due to genetics or whatever variables are different, if any.
trichomes are great, but do not always amber up. There is a common misunderstanding that gets repeated. Amber is absolutely not "peak" thc-a. It is visible degradation. THC and it's other forms that we care more about are mostly colorless and odorless... attributing either to it is madness. Cloudy is beginning of degradation (to cbd? whatever it is, it is useless). the reason we want 'some' amber is because of the staggering growth of calyxes that make up the buds.. some are older, some are younger and the "optimal" point likely has some of each - clear, cloudy and amber.
Now, earlier harvests supposedly have less couch lock. this is directly related to the degredation that is occurring. If you don't want the sleepy groggy high, you want an earlier harvest in the window... This is where using 3 or more visual clues also helps... you can do early - middle - late and test it for yourself... strain may make differences too, but relative to any one plant these correlations are sound.
Trichomes
Pistils
Density of buds.
use all three. you can good for the "window" as well as i can guess at them... 10-70% amber etc... pistils at least 1/2 coloration.. i prefer to let them recoil and that is usually overlaps with early window trichome coloration. I prefer earlier window harvests. I've done longer with same exact clones and dd not like it when it made me sleepy. YMMV... try it all before you assume it is the same for you... like alcohol... it makes some energetic it makes others the opposite.. same chemical impacting both outcomes, but different brains.