I am just going to dump here a reply I gave on another forum.
TL: DR – Nothing really changed in my eyes…
So I read the scientific article and haven't read any of the news articles on the subject. The paper studied 1964 horses from the Late Roman period up to the Post-Medieval period, the horses were found on 171 sites. The article does say warhorses were smaller than in "popular perception". Not 17 and 18hh but maximumly 15 to 16hh, which would have been tall for a Medieval person. But it does not explicitly state that the warhorses of the Medieval period were ponies.
A large part of the discussion goes on about the difficulties of the research on (Medieval) warhorses. How it is difficult to determine the activity of a horse from a full skeleton and more research is necessary to determine the activity of horses through measuring of bones. It also notes that most specimens that have been used for this paper were single bones (depending on the bone, a bone can say a lot. If you have a bunch of front legs (metacarpus) of B.tauros, you can determine for example which are bulls, cows and oxen).
It also discusses the context. When you find a horse in a military context, like a castle, it does not necessarily mean that the horse in question is a warhorse. There is only one mass grave with horses associated with a battlefield in England. It contained dismembered horses, which is not odd for the Medieval period as horses, even warhorses, were post-mortem processed for valuable goods. (I am familiar with horse burials of the Merovingian period, but these were normal cemeteries).
The article also mentions how warhorses would only form a small selection of horses in the Medieval period.
(Ameen et al. 2021, p.1250, fig.2)
Ameen, Carly, et al. "In search of the ‘great horse': A zooarchaeological assessment of horses from England (AD 300–1650)." International Journal of Osteoarchaeology 31.6 (2021): 1247-1257.
So with that out of the way, what is my interpretation. For starters, what is the big deal? In my eyes, it is more a confirmation of expectation. Medieval warhorses were smaller than modern-day horses.
Because the research mostly looked at induvial bones, rather than individual animals, I wonder about the estimated individuals overall. To me, it seems the actual number of individuals could be lower than the 1964 specimen but that's the challenge with zooarchaeology. Thoughts there are some methods to give a suggestion it was not discussed here. It did discuss however the possibility of some misidentifying with donkeys, mules and hinnies.
It should also be noted that the sites came all from the UK, mostly from South and central England. Thus it does not have to reflect the regions on the continent. Consider cultural and biological influences.
It could also be interesting to measure surviving horse armour, as by definition it would tell us something about warhorses.
Regardless I recommend everyone to read the scientific article for themselves and draw up their own conclusions. The article should be freely accessible through Google Scholar.