There are two approaches to estimate tasks
When you estimate, it is OK to trust your gut or intuition. It should rarely be your primary basis for an estimate, but an estimate feels wrong, it probably is. And although gut feel / intuition can play a small role in estimating some estimates, there is no room for wild guesses that aren’t based on anything.
A better approach than gut feel is estimating by analogy. This refers to estimating something by comparing to some other, similar thing with which you are more familiar.
A good technique when estimating is to triangulate estimates. This refers to comparing the item being estimated to two, or occasionally more, items that have already been estimated. These comparisons often to one item larger and one smaller than the item being estimated can help indicate if an item is about to be given a good estimate.
Triangulating is the best way to ensure that items are estimated consistently. It refers to comparing an item being estimated to two or more other items. Usually the item will be compared to one larger and one smaller item, but that is not required. Sometimes the item being estimated will be compared to an item of the same size and one other that is either smaller or larger.
Because triangulating can become a natural part of the discussion while estimating, it should be done on all or nearly all product backlog items. It is not something you use on only the most important items.
The last word about estimating that I want to say is be careful when disaggregating stories into smaller tasks. A little bit effort goes a long way, but more effort only helps a little.
Round estimates up when needed. Work expands when looked at in detail.