It seems the differences are coming from the Clarke tanks and not the TW's. Seems like a manufacturing process(es) are throwing off the end product. Are there multiple molds Clarke is using where one produces a good product and one does not? Is there a specific temperature that the newly molded tanks must reach before being removed from the molds? What happens if that temperature isn't reached, are the newly produced tanks distorting while cooling? Are temperatures at which the tank is stored/shipped/etc. coming into effect at altering the shape of the tank? Are the molds properly designed to deal with thermal expansion? I would hope so, but maybe a newbie designed it incorrectly and it got through the checks and balances without getting caught.
Molds aren't cheap. They may acknowledge a problem internally at Clarke but understand it's too costly to redesign/cut new molds, and that acknowledgement will never come out publicly, but is something we can only hypothesize about.
My limited knowledge says no matter what stock metal tank you find (an '87 or a '95 or an '09) they will all bolt up exactly the same...which should prove that it's not a TW issue, but a Clarke issue.
My 2 cents