It is common to find the Sanskrit term tridaśāḥ translated as “thirty Gods”. This is presumably based on Monier-Williams’ dictionary entry:
m. pl. (cf. Pāṇ. 2-2, 25; v, 4, 73; vi, 3, 48 Kāś. and <dvid->) the 3 X 10 (in round number for 3 X 11) deities (12 Ādityas, 8 Vasus, 11 Rudras, and 2 Aśvins; cf. RV. ix, 92, 24) MBh. &c.; du. the Aśvins, iii, 10345;
This entry does not dispute the fact that there are actually thirty-three Gods and that when authors intend to name them by a numerical appellation we commonly find: trayastriṃśāḥ. So how are we to interpret the form?
1. It is simply a metrically more convenient form for trayastriṃśāḥ. For instance, in the Rāmāyaṇa they often follow each other fairly closely, and so must mean the same thing. This does not really address the issue at all: of course a poet would use the forms in metrically correct positions.
2. What else could it mean.
I believe that it is of course entirely plausible that there are cases where tridaśāḥ can and indeed should be interpreted as outlined above. This is borne out even by some commentators on the Amarakoṣa such as Kṣīrasvāmin (and Mallinātha on Kumārasambhava 3.1: trir āvṛttā daśa parimāṇam eṣām iti tridaśān devān. Incidentally, the earlier commentator Vallabhadeva glosses this passage simply by devān, which I would take as an endorsement of the second interpretation given below). It seems equally clear to me that this is hardly a universal, or even a very common view in the Sanskrit scholastic tradition. This is a purely subjective opinion based on what little I have read in commentatorial literature.
The other, to my mind more common, interpretation of the compound is as a numerical bahuvrīhi:
tisro daśā yeṣāṃ bālyayauvanamadhyatvakṛtās te tridaśā = devāḥ,
“Those who have three life stages, namely infancy, youth and middle age [but no old age] are the Gods.”
This is in fact the analysis provided by the Amarakoṣaṭīkā 1.1.7 of Jātarūpa, the earliest surviving commentary on the Amarakoṣa (composed end of 10th cent CE, see Mahes Raj Pant 2000:283–308). Kṣīrasvāmin himself knew this interpretation and attributed it to the “Gauḍa,” confirming that Jātarūpa worked in Eastern India, while Jātarūpa does not mention the (later?) “rounded number” theory. Nor do many later commentators from South India make any mention of it (such as Liṅgayasūri in his Amarapadavivṛti: tridaśaparimāṇaṃ vayo yeṣām iti tridaśāḥ).
Even if we can accept that tridaśāḥ might for some commentators have meant “thirty” for “thirty-three” (though I believe the absence of attested variation such as triṃśa etc. for trayastriṃśa in this sense speaks against this being the original sense) it is still inadvisable to translate this as “thirty Gods,” for there is no way a reader can then guess that the number is actually thirty-three—there being no well established idiom in English of thirty-three Gods to which the sense might readily shift.