I voted hope and guess Civ VI. The game + two expansions business model seems strong. I would suspect that only a third of people who buy the game buy the first expansion, and that only 2/3 of those who buy the first expansion buy the second. You're whittling down to your hardcore fans. You need a full-fledged new game every once in a while to create new buzz about the franchise, and again, get the big initial customer base.
It's possible this is the typical model, however in the DLC era there's more demand for 'intervening' content to keep the game fresh before the release of the next sequel. What's more, Civ IV had three 'expansions' including Colonization (which in modern parlance would qualify as a "standalone expansion" like Total War's Fall of the Samurai, a spinoff built from the same game engine but not compatible with the main game).
It is also possible that BNW changed the maths involved. This expansion was very well-received and the fact that Civ V is still among the five most-played games on Steam three years after release suggests that it's not losing players - sure Steam stats don't track how many of those players are buying the expansions, but Steam undoubtedly collects that data and can give the developers a good idea of whether a third expansion is a good idea.
As for the "Bison" project, are they likely to be secretive or giving code names for something like a civ DLC alone?
Besides, the even numbered games are probably a little quicker to get to market. Civ VI will be able to use most of the same mechanics (hexes, 1upt) that CivV innovated. Naturally we all hope that they'll program the AI to fight better in those mechanics, but in any case, they're not designing them from scratch.
There's no fixed rule that Civ games have to have radical changes for odd-numbered incarnations. A lot of Civ V's vanilla mechanics were unpopular (indeed it wasn't until BNW hit that the "Civ V Rants Thread" finally vanished from the first page of this forum); the expansions (particularly BNW) were more comprehensive in their changes than those for previous Civ games but still worked within the confines of a system designed by a different team.
It's very possible that, if the BNW team is put in charge of developing Civ VI (which I imagine most people would agree would be the optimal decision), that they'd want to go back and 'reimagine' core mechanics they weren't able to change in this version of the game, either because they're hard-coded into the engine or because they couldn't find a way to fit them into the themes they'd set themselves for the expansions. For example, take the combat system: what we have at the moment is a "workaround" for an AI that struggles with 1UPT, in making a game that promotes more peaceful play where the AI is able to present more of a challenge. A new Civ game will want to go back to basics and determine whether new AI coding can better-implement 1UPT (and it's not that bad now), whether maps for a given number of civs need to be made larger to limit 'traffic jams', or whether some form of stacking can be added, with attendant changes needed to production rates, unit type balancing and targeting (if you were to stack 1 melee on top of 1 ranged, it's not satisfactory to have the current "only the unit on top can be targeted" system).
I think there's scope for a third (and probably only a third) expansion in what's 'left out' of Civ V so far, but I suspect that, following Civ IV precedent, there'll be a spin-off game to fill the gap before Civ VI - so I voted for "hope expansion/expect spinoff".