"This horse is mine."
Translation:Αυτό είναι το άλογό μου.
Modern Greek can have two stressed syllables per word if the word is followed by an enclitic - a small word that has no word stress of its own.
Adding the unstressed enclitic sort of turns the word into a new, longer word -- and if that longer word has the stress further back than the third syllable from the end, you need to add another stress.
That is what happens here -- "το άλογομου" (as it would be pronounced, though not written) has the stress too far back, so it has to be "το άλογόμου", which is then written "το άλογό μου".
Compare some long English words such as "idéntificátion" from "idéntify" -- where "idéntification" would have the stress really far back from the end. (Plus -átion is usually stressed anyway.)
That's right - το άλογο μου is incorrect.
When composing sentences for the course, you do so by selecting words that have already been taught in past lessons or that are part of the current lesson, so it's possible to select "το + άλογο + μου", but "άλογό" isn't in the system as a word to be taught so it's not easily possible to add it to the sentence properly.
I think it might be possible to fix this by using "Forms" (where one "vocabulary word" is actually a list of related forms, which could be for things such as singular+plural, but could probably also work for single/double accent) but I'm not sure how much effort it would be to convert all the nouns like that now.... and I believe that in the beta phase, all Greek sentences are frozen anyway, so they can only add alternative translations in Greek or English but can't change the main sentences. So any fixes would have to wait for a possible second tree anyway.
So as I understand it, the plan right now is to add the correct versions as alternative translations to all English-to-Greek exercises, but they can't do much about the existing incorrect Greek-to-English ones.