OK I accept the logic... but why then do we just accept the surface wind conditions? The wind has a far, far greater effect than anti-ice but we don't have to consider wind effects above the surface. At what height then do you decide that the cloud no longer has any effect? You certainly wouldn't select anti-ice on before TO if there was cloud at 1000ft (or 1500ft or 2000ft), that would be selected in the climb, but the debate in my company is whether the ice correction should be applied if you feel that ice would be selected at some stage during the climb out.
(Obviously if you elect to use anti-ice on the TO roll then you would have to make a perf correction but nobody is going to pre select anti-ice for a dry RW because there is cloud at 1000ft)
Would love to find a ref that confirms it either way.