This is amazing - I wonder if some luck was involved. I’m sure they were deliberately trying to make strong/lasting concrete but they couldn’t have tested it over hundreds of years, let alone 1000s. Is there a degree of chance to it being this effective?
Everything involves luck, but until very recently (past 100 years) mortar maker was an expert occupation with thousands of years of accumulated art and science. These days most professional concrete and foundation contractors don't have a clue about mixing or pouring anything other than portland cement and whatever is most popular at Home Depot.
Who knows how much knowledge has been lost, but my guess is that the Romans developed their processes much more deliberately than haphazardly. You don't have to understand the 21st century chemistry of why something works in order to build a systematic and even scientific methodology. It was part and parcel of a mortar makers job to understand how to make use of regional materials and handle local environmental conditions.
This seems like it's a great example of survivorship bias: the Romans probably made lots of concrete that didn't survive the elements, but there was some that the Romans did get lucky with. It's also possible that we don't see a similar thing today because our manufacturing processes are so much more standardized (and critically, they are standardized around a process that is apparently less robust.)
There are literally thousands of types of concrete you can buy today, and even now occasionally you get a 'bad batch' that doesn't test as strong as it's supposed to be. That indicates there must be quite a bit of per batch variation too.
Overall, considering we build far more things per day than the Romans did, I reckon there's a good chance more of our buildings are around in 2000 years than the Romans have left now
The point is not much about strength (and measured strength according to norms) but rather about durability and strength over time.
With "modern" cement (Portland) it is relatively easy to get high strength at 4 weeks/28 days, while (good ol') pozzolanic cement the reaction/hardening is much slower but over time (and with adequate humidity) it can reach and beat the "better" portland cement.
Almost nothing we build today will be there in 2000 years, but not because of the concrete itself, but because of the reinforcing steel we use (and because of the different use of concrete in much slimmer structures).
Anecdata: in the '80's/'90's I was working in a large tunnel project and we used in tunnels pozzolanic cement (not reinforced) for the lower part of the lining, and it was not easy to get the "right" (according to norms) 28 days strength (250 Kg/cm^2 at the time), but, once the tunnel was finished (roughly 3-5 years laters) and we did further testing of the structures, we found that it reached strength of the order of magnitude of 500-600 Kg/cm^2 whilst the corresponding Portland based concrete, targeted at the same 250 Kg/cm^2, reached "only" 300-350 Kg/cm^2.
Yes, though the other issues remain and stainless is very different from "ethernal" (much longer lasting, yes, 2000 years I doubt), besides stainless steel (which is very expensive, 4 to 6 times common reinforcing steel as an order of magnitude) also zinc or resin coated steel is used.
Concrete is only a good material to resist compression forces - generally speaking - it was used by Romans only in structures that were exclusively (or almost exclusively) subject to compression (arches).
The revolution that reinforced concrete made was about combining two materials (steel very apt to resist tensile forces and concrete very apt to resist compression) with very similar other characteristics (thermal expansion) and compatible between them, if you want it is one of the first examples of composite material, to obtain something that could be used in structures subject to tensile and compression (besides shear) forces.
Our recently (last 100 years or so) reinforced concrete structures are very lean and elastic, which implies that they move and crack.
All the research on new concrete is about making mixtures where cracks are reduced to the minimum, as before or later through these (micro) cracks air and water penetrate, oxidizing (or rusting) the steel.
As well in the years the norms about the cover (i.e. the minimal distance from steel to the outside) has been increased (it depends on countries and types of structure but 30 years ago 2 or 2.5 cm were common, nowadays 4 or 5 cm are common).
