Your screenshot shows the bot Nora, which is rated 2200. Does this mean that your comment about Elo ratings only applies to the higher-rated bots? I'm asking because I beat Wendy (rated 1500) in a rapid game (10 minutes each) and did not have the impression that her play was anything near a 1500 rating. Without the rating indication, I would have assumed a rating of at most 900.
And I don't think my win today against Wendy was a fluke; it's not my first win against that bot.
I can't comment on the time-management aspect of the question, since I haven't played any other slow time-controlled games against bots.
I think you nailed it with the bots ELO. So far, their range of levels and the types of mistakes they make feel very well tuned, at least from my perspective.
I do have one question: have the bots been upgraded to also mimic time management based on their corresponding ELO? For example, the time they spend per move, how they behave under time pressure when the clock is low, occasionally losing on time depending on their level, making flagging attempts, and so on. You get what I mean. Since you added bullet, blitz, and rapid prominently to the bot options, I’m curious whether the bots were improved in this way to make them even more realistic.
CHEERS!