This test is funny. It confirms what I have posted. Thermal paste has only one function, to cover micro scratches. Therefore the viscosity/texture plays a role, nothing else. BUT: What also plays a role is the chemical (and physical) stability during time. Honey and ketchup have got a lot of sugar which decomposes during 'use'... so viscosity and texture change very quickly. Thermal paste remains both original as long as possible. The thermoconductivity of the paste itself plays no role. So adding 'silver'/ advertise it is made to fool people.
I think the main problem is what happens over time. Any paste is probably a good conductor of heat. But when it dries into a dry plaster, I would think that would be more of an insulator. You want something that stays pliable over time. It's probably wise to reapply thermal compound every couple of years, especially if you overclock and generate extra heat that may tend to dry out the compound.
As Yen has pointed out, thermal compound sits only in the voids of the mating surfaces. If the two surfaces are flat and smooth, there is no place for the compound to reside. Ever hear of Johansson blocks? I ran into these decades ago when taking a machinist class. They are precision machined to be flat and accurate to one millionth of an inch (at a given temperature, as thermal expansion changes their dimensions). Place one Jo block on another and you can not pull them apart - they just about become a solid piece of material. The only way to separate them is to slide them apart. (I see these days they are also made of ceramic) Obviously, no one makes a heat sink that is flat and smooth to a millionth of an inch, but chances are any thermal compound is only conducting heat through 1-3% of the surface area, the rest being metal to metal or ceramic) contact.
Only for those who have no clue about physics and thermal flow. . They are victims of a hype. When you have a well made OC cooler it is planar enough already so the paste plays almost no role.... Other things such as proper air flow within the case are more important. I build my own PCs and some for friends (I used Noctua). The CPU cooler itself and the additional fans' position (air flow) are significant, the brand of the paste is not. Too much paste is also contra productive. These are my own practical experiences. The issue comes mostly from the CPU package surface. Heat sinks' is actually flat and smooth...(better)
I have come across an Acer system that switches off if you apply thermal compound and if you wipe it clean(without) thermal paste it works just fine. Not all heatsinks need it.
This can happen for a number of reasons. Thermal compounds differ widely in their electrical conductivity and capacitance , and if excessive quantities are applied there can be "spill-over " onto the pins of the chip or the circuit tracks. -