This etymology as the competing one parses the word as a compound with the preverb ἀνα-, under the syncopated form ἀν- / ἀμ- which was familiar to Greek scholars because it is well known in Homer and in several dialects. The second element is supposed to be πηλός "mud", which as a secondary meaning refers to dregs, the deposit left by wine. This implies an alternation between [ĕ] and [ē] of the τίθημι / τίθεμεν type. The interesting thing about it is that πηλός itself is assumed to be derived from πάλλειν, which implies a different alternation between [ă] and [ē] of the ἵστημι / ἵσταμεν type: Orion says elsewhere that this etymology, too, comes from Herodian (see πηλός / πάλλω). It was not a problem for Herodian to combine both in his derivative pattern. We modern know that those alternations show up in different roots and cannot be combined in the same root. But the notion of root was alien to Greek scholars (except maybe for Philoxenus who comes as close to it as he could), and each step, πάλλω / πηλός, and then πηλός / ἄμπελος, could in fact be backed by similar cases. The question of the compatibility of the two was not asked