Time to look at those two D/L revisions against the Jayadevan system:
Northants were so utterly buggered that VJD would also have set very little - 25 in 5 overs. Northants were 31/4 off 8.3 overs, having used 50% of their total resource (43% overs, 40% wickets). VJD tells us that the target runs in an 8.3 over chase is 56.1%, so Northants' effective 8.3 over score = 56.1/50 * 31 = 34.782. Gloucester have 5/8.3 = 10/17 = 58.8% of these overs remaining; this means they should be expected to score 71.3% of their effective target, so 0.713 * 34.782 = 25 to win. It'll be tough to argue any higher target with any logic that you can extrapolate to another score - teams are not usually 31/4 in 8.3 overs.
At 104/0 off 9 overs, however, Sussex have used 45% overs, losing 0% wickets, which the VJD table tells us is 36.3% of the final total. Target runs in a 9 over chase is 58.6%, giving an effective 9 over score of 58.6/36.3 * 104 = 167.89. Surrey have 5/9 = 55.5% of that chase; meaning they should be expected to score 68.5% of that total. 0.685 * 167.89 = 116 to win in five overs. If I can read, because it's bloody confusing and I've spent three days in a tent in Wales.
Draw your own conclusions...
S Bhogle on VJD
VJD Tables