I wouldn't take the game approach.
I think I'd focus on the different ways appraisers and (many) brokers define value. It's their job to buy and sell, so it only makes sense that in that role they need to find the 1 willing buyer and 1 willing seller.
Our role is different. The deal at hand may or may not reflect the market independent of what this particular buyer and seller think. Some deals close high and some close low so that's why the contract price doesn't drive the outcome. We're looking for the central trend among many, not the highest or lowest "we can support".
We need to be mindful of not expecting brokers to think like appraisers, and not expecting appraisers to think like brokers. Both will be counterproductive to our respective roles. You gotta be you and I gotta be me.
We don't search for comps by price because "price" is the answer to the question, which means it isn't a search parameter. We're looking for proximate, similar age/size and similar lots. We may only put 4 or 5 sales in a report but in order to get to that 4 or 5 we will also be looking at many more. So in that respect the 4 or 5 are the tip of a large iceberg of data we're looking at. That's why - even if the broker thinks they're more similar - swapping out 1 or 2 of the comparables won't do anything to to alter the opinion of an appraiser who is working to specs.
Oh, and it never hurts to hand an appraiser a list of whatever features the seller or broker thinks are important, there's no reason to leave it to chance that the appraiser will notice everything they need to notice.