Yes, your analysis is correct. In fact the し can drop even lower than the う before it if you choose to really enunciate it. This sort of splitting is fairly common, for example with the prefix 非 or the prefix 被.
I consider it wrong to read that word in heiban, it'd sound like 調子・全的(??) or something like that.
However, there are sometimes cases where words get reanalyzed, like
無期懲役（むきちょ↓うえき） → 無期懲役（む↓き・ちょうえき）
or
故事成語（こ↓じ・せいご）→ 故事成語（こじせ↓いご）
This tends to happen as the compound becomes more common and non-experts come to say it more IMO. I don't think there is any chance of this happening with 超自然的 because it isn't a single concept, though, 的 is a suffix.
There are also other cases like 興味本位（きょうみほ↓んい） where it's not valid to pronounced split (×きょ↓うみ・ほ↓んい), because 本位 is not a valid concept on its own, it only serves as a suffix so it makes no sense to have it split.
大辞林 tends to list such split pronunciations as number-number, like 切磋琢磨:
せっさたくま ［1］━［1］ 【切磋▼琢▼磨】
Which indicates it's a two-part word which drops after the first mora of each part. However, it doesn't indicate where the break is so you sort of need to know that yourself.
In addition, in the case the second part of the word is heiban, it seems to omit the -[0] (like with 春夏秋冬), which is not good IMO because [1]-[0] does permit a slightly different pronunciation from just [1].
For these reasons I recommend using the NHK accent dictionary, which does show where things split and their respective downstep locations. Of course the word you ask about isn't a dictionary word so it wouldn't be helpful there.
