1998, Jazz vs Pacers

[SportsGuru08]

Who wins this hypothetical Finals series?

[Special_Puppy]

Definitely the Jazz. Ended up with a much higher blended elo
https://fivethirtyeight.com/features/the-warriors-are-in-the-goat-debate-but-they-blew-their-chance-to-end-it/

[Owly]

Definitely the Jazz. Ended up with a much higher blended elo
https://fivethirtyeight.com/features/the-warriors-are-in-the-goat-debate-but-they-blew-their-chance-to-end-it/

(FWIW, my instinct is a playoff rotation Jazz is better, Jazz full-strength would be undersold by season long numbers with Stockton missing time ... more conceptually though ...)

That is a (semi-)rolling number though right?

So Jazz ended the previous season at 1748, Pacers at 1516 ...
and iirc they regress it back towards average a bit but that's a factor?
And so the 97 team where Smits, Jackson and McKey missed significant time is a bit of a drag. Some roster changes Mullin (old) in for Dampier (young), Duane Ferrell no longer in the rotation.
(As before in season measure may not fully capture Jazz with Stockton strength).

Also I'd need to get more familiar with ELO but instinctively
- wouldn't average daily ... looking at two seemingly identical teams (say both .500, both 0 SRS - say both miss the playoffs for simplicity) favor the team who was good early over the one that was good later
- wouldn't peak favor the early team too.
- "Daily", too? Is the time in between games really supposed to matter?
- At a glance they ignore very early season numbers for this purpose, which acknowledges and somewhat mitigates ... but not fully. And I get the idea of someone using, say, surrounding seasons as something of a prior - I'm not saying no data from outside the season in question - but for assessing a season this all seems quite messy if I've understood it correctly.