I tend to be overly excited and thus think that I can finish something that takes a week in a day. Then when I give myself more time for something I finish it more quickly...
The key is to excercise the estimation muscle.
Keep logs about your estimations and how long it actually took.
Then think through periodically what caused the big differences.
This way in the long run you’ll be able to precisely estimate tasks you’ve sort of already done in the past, and can estimate a reasonable low/high time range to finish tasks which you’ve never done before.
> Typically the goals I choose are things that might take one to two hours of focused effort.
Very few of us can accurately estimate what will take a day, but an hour is much more manageable.
In addition, on the few occasions where I didn't manage to hit my goal by the end of the day (maybe three times in the last month), I'd take the time to figure out why that was and think of how I could have reframed the goal in a way that would have made it achievable. Usually when I missed a goal it was because it was either not specific enough or it was dependent on something else. For example, I might have had "Push XYZ patch", when it should have been "Make bugfixes to XYZ patch and push if able", since there's always the possibility that the latest fixes introduce an unexpected problem or that the reviewer won't get to the code review that day.