The "motion lighting" feature they (the people who did the test) specifically mention: upon enabling it & resuming playback of content (such as a game in a dark locale but with a lot of movement), the backlight will dim to sort of match the content. Even if the content brightens up, the backlight will not immediately max out. The in-menu description does a poor job of describing how the feature will work, but it does specifically say it's to reduce power consumption. Disable it, and the backlight returns to its normal setting. This shouldn't be a surprise.
Another feature (LED Motion Plus) allows the backlight to dim slightly in an attempt to reduce flicker (have the LEDs refresh more frequently but put out less overall light). This would probably reduce power consumption, too, but is not the focus of the feature.
Let us not forget that some of these TV sets feature ambient light sensors, and in an area of low lighting with this feature enabled the TV could easily be set to allow the backlight to go to 0 automatically. Put into a bright room, the power consumption would change.
Even the default "Dynamic" preset (not marked with an EnergyStar logo, unlike the "Standard" preset) requires some use of these features to actually have bright brights and dark darks. It doesn't force the backlight to be at 100% brightness 100% of the time, to my knowledge.
I would love to know how the TV would detect it's being tested. Unlike with the phones (which could actively know the name of the processes running on the phone & detect the load patterns as they taxed the hardware), the TVs should effectively be unaware of where the source material being displayed is coming from.
If you're using the right source material, it shouldn't be a surprise that this power conservation happens, as this is similar to dynamic contrast & dynamic backlighting (including local LED dimming) which helps sets achieve deep blacks while maintaining brilliant light colors, just with a slightly different purpose in mind.
There are discussions on the Internet where users ask about these features. It takes 0 effort to figure out what they will do. I don't see how this is a cheat or an attempt to fool anyone; it sounds like someone refused to actually go through all the settings & make note of what each feature does BEFORE testing. If you make your testing methodology public, you could present every setting you used on the TV and verify that others who use those settings get similar results. This is why getting a good testing methodology is important; you need to be able to hit edge cases where things may be significantly different & account for those. Regardless, everyone should be able to hit the power targets mentioned by Samsung in its marketing & specifications.