Token order matters more than weighting syntax in most modern models — a demo
Technique flair. Everyone obsesses over (word:1.3) weighting but in a lot of current models, where a concept SITS in the prompt does more than any numeric weight.
Front-load what matters. The subject and the single most important attribute should be in the first ~15 words. Models attend more strongly to the opening; stuff buried at the end gets diluted or dropped, especially as prompts get long.
Quick test anyone can run: take your prompt, move your key attribute from the end to the front, change nothing else. In most modern models the front version wins noticeably. The 'weight' you wanted was positional all along.
Weighting syntax still matters where it's supported, but position is the free lunch everyone ignores.