A tactical tilt might take equities from a 60% policy weight to 50% because valuations look stretched, or add 5% to commodities on an inflation view. The size of the permitted deviation is usually capped in advance, for example plus or minus 10 percentage points per asset class.
Tactical allocation is an active bet and should be judged like one. If the tilt is 10 points of equity and equities underperform bonds by 4% over the period, the tilt adds 0.10 x 4% = 0.4% to return. That is the payoff for being right; being wrong costs the same. Few tactical programmes clear their own costs over a full cycle.
Keep a record of every tilt, its thesis, and its exit condition, the same way you would keep a trading-journal for individual trades. Without that record it is impossible to tell skill from a run of luck.
Related: strategic-asset-allocation, asset-allocation, dynamic-asset-allocation, sector-rotation, risk-on-risk-off