Originally Posted by
SFBayGuy
No, no, no. Let's use some (made-up) real numbers in lieu of the unknowns we've been working with. Say that each chain has 10 properties with 100 rooms each, 2% of which are "upgrade" rooms. So that's 1000 rooms per chain and 20 upgrade rooms per chain. In the combined entity there will be 2000 rooms and 40 upgrade rooms. Simple addition.
Now suppose each chain has 1000 program members, 10% of which are elite, so 100 elites per chain. Here it's not quite as simple, but it's still high school math. For the sake of this example we'll say that 50% of the members in MR are also members in SPG and 25% of the elites are elite in both, so there will be 1500 (500both+500MR+500SPG) members and 175 (25both+75MR+75SPG) elites in the combined program.
Bottom line: Before: 200 elites vying for 40 upgrade rooms. After: 175 elites vying for the same 40 rooms.
The above said, yes, there are now 75% more elites in the pool for any single hotel's upgrade rooms on any given night, but it is faulty logic to generalize that added 75% across the entire chain because if more elites show up at one hotel that means fewer showed up at the others.
From a probability viewpoint MR, SPG, and "both" members are largely biased (not randomly selected) populations. In other words the groups are self selected ... SPG members have largely found that SPG hotels have locations and service levels that meet their needs better than Marriott. Similarly for MR members there is a bias towards Marriott properties. There will also be a component of perceived program quality in the selection. Those in both are also there for reasons likely driven by property locations and service levels.
So the largest change in upgrade probabilities logically will be from the "both category" who will, in the combined program, tend to move to a higher status level (assuming they did not have so many stays per year that they were maxed out in both programs). This group already stays at both Starwood and Marriot properties and are most likely to continue to do so.
As a result the most important bit of information to estimate future upgrade probability changes in a particular city would be to know the number of members of the "both" category that on the average stay in your target city. Not sure even Marriot would have those numbers...