My understanding of story estimation has been that one should estimate the size of a story as it would be for an imaginary, average developer — a bit like the "reasonable bystander" concept in law. That is, you should not estimate the story's size assuming you have to do it.
To give an example: in my previous job I was part of a team where I was far and away the most confident Ruby developer. My teammates would routinely estimate Ruby-related stories far bigger than I would, with arguments like, "Well I don't know how X works in Ruby, so this would take me ages to do."
My argument against this comes from the fact that sprint planning is where the team's capacity comes into play. That is the correct forum to say, "Our capacity this sprint will be slightly lower than usual because the majority of the tasks are Ruby-based, and we only have one strong Ruby developer." Factoring this in during estimation would double up this aspect.
I'd appreciate any authoritative references in answers, but simple opinions would be great too.
Best Answer
Story points are a relative estimation. So twice the points means twice the level of effort. Relative estimates are less subject to skill level variations. The question is not so much how much time you would take for 1 point, but that 2 points require 2 times more potential effort. Skill level could matter more if you would take ideal days instead of story points, because you assume an individual productivity level.
Relative estimates are more robust. In addition, the story point evaluation should not be performed by an individual, but result from a collective team effort. For less complex stories, there's usually a quick agreement. For more challenging stories, the team will come with a result to which most of the members will agree, and therefore implicitly take into account the collective skill level of the team.
Finally, story evaluation is performed at a moment when assignments within the team are not necessarily already decided. This is one more argument not to take into account individual skill level. For sprint planning, you'll take the story point capacity of the team, which is a figure that will evolve based on actual performance figures, so that it will self-adjust to the global skill level of your team.
In conclusion, the individual ability should not be taken into account for the estimate. But even if it would be done, due to the collective estimates, and the robustness of the relative approach, it wouldn't matter so much.