Remember, the odds are looking at the overall picture and playing out everyone's schedules 5,000 times simulated to create a multitude of scenarios that eventually come to a % of times that Portland is the 8-seed vs. percentage of times that Memphis is the 8-seed, etc. If Portland is playing at home to the Nets, nothing really changed by the one win, they were projected to win like 4,995 of those 5,000 games anyway. So losing that game would cause them to plummet, but winning doesn't mean much. Similarly, Memphis losing to the 3rd best team in the East doesn't impact their % much since they were expected to lose.
As the weeks go on and teams like New Orleans and Denver pull 1-2 games further ahead of Portland (with less remaining games to be played), their odds amplify even higher since there are fewer games for them to lose ground and for Portland to make up those games. So those teams all get higher percentages and Memphis and Portland get fewer of those chances of those teams plummeting.
Third, there is also the possibility that future teams that Portland play on their schedule got better in the last week and it changes the odds Portland wins those games in the 5,000 scenarios. So if Portland plays San Antonio and the Lakers like 6 more times and Memphis only plays them 4 times, meanwhile LA and San Antonio are on fire in the last week, it possibly lowers Portlan's win total by 0.3 compared to Memphis' 0.2, in which that 0.1 difference causes Memphis to get the 8-seed in a few thousand more combined scenarios (out of the hundreds of thousands of combinations) than they would have prior to LA and San Antonio having this hot streak. Just an example of how the play of other teams affect those teams hovering around the 8-seed.
