This is an interesting question and I think extends beyond just NY and extends to races like WA-Sen and CO-Sen.
I think if you actually want to be tactical with this strategy, you'd want to release your polls slightly later than Trafalgar did. The influx of close R leaning NY polls started around early September which gave Hochul some time to respond and probably solidified partisanship a bit just now that many saw the race on the map.
I think the bigger way in which polling sort of screwed things up was in the House where Democrats triaged a lot of House races that in hindsight they really shouldn't have. At least Rs wasted their money on WA and CO-Sen lol.
Arizona's a good example of this. In normal circumstances, there's no way that O'Halleran would be favored in Arizona 1, especially with it being redrawn to be a bit red, but Democrats really favored protecting incumbents above anything else, so money was spent there. There was also some internal polling that showed O'Halleran to be in much better shape, that turned out to be inaccurate.
Turns out that the 2nd and 5th were way more competitive, and with more money and hindsight, they could have been Democrat victories. Alas!