It depends on quite a number of factors.

1. How far back in time are the events being recalled?- A "traditon" can be defined as a narrative of events passed down one or more generations. That is, the narrator is recounting something they have been told, not something experienced first-hand.

It seems obvious to state that recent traditions are more likely to have more detail than older traditions. e.g. when an informant tells a family story about their grandfather as a young man, that informant can draw on personal experience to fill in detail about grandfathers personality. Family knowledge could also let the informant talk somewhat about aunts, uncles, cousins and neighbors.

But, if our informant were to talk about their Great-great-grandfather, it's much less likely they would be able to speak about that generation's relatives, neighbors, or the ancestors personality.

So, as oral traditions become further removed from the events they detail, you tend to have fewer minor characters and only very noteworthy individuals tend to remain. Similarly, very noteworthy events like fires, droughts, wars, and weddings are remembered, but more day-to-day events tend to fall out of the narrative.

Also, individuals personalities or backgrounds can be transformed into cliches. For example, in Central Africa, histories of royal lineages almost always start with a founder figure who is a hunter/warrior who expands the kingdom. He is usually succeeded by a king who organizes the kingdom, acts as lawgiver, and establishes the forms of royal administration.

2. How many informants do you have? or, how many versions of the tradition do you have?

As oral traditions are passed down, they are subjected to societal forces. Traditions can become weapons to justify an existing social order. Likewise, with changes in a social order, traditions can be reinterpreted in the light of present circumstances.

So, to try and mitigate that, it is important to try and find multiple informants about a narrative, ideally from different families/villages/societies. Having multiple versions of the myth, you can compare them against each other and see where the variations in interpretations are. Also, some versions may recall information not present in other versions of the tradition.

That's not to say the additional information will necessarily be true, but it will be more information to analyze, and compare against other sources of data (e.g. Archaeological, linguistic).

3. Are there mnemonic devices in the way the tradition is told?

For instance, Mandinka griots telling the Epic of Sunjata sing the sing in a stylized way, with specific meter, while playing the Kora. So, the meter within the song helps to organize the structure of the story and prevent the loss of details. In that way, stories can be standardized and passed down generations with limited loss of detail.

4. what was the structure of the interview when the tradition was written down?

This can strongly influence what information is shared. Is the interviewer focusing specific questions? Are they allowing the informant time and room to digress?

Does the informant see the interviewer as in insider or as an outsider? That is a big one.

5. are we fully understanding allegory?

Sometimes, traditions can convey information in "coded language" that would be understood to someone belonging to the group, but not obvious to outsiders. Take for example this passage from Central Africa to 1870: Zambezia, Zaire and the South Atlantic

The Lunda were governed by elders known as tubungu who com- manded respect by their seniority, their experience, and above all their spiritual powers. These elders were known not by their personal names, but by the titles of the offices which they held. Relations between the Lunda title-holders were governed by fictitious geneologies. One title would be deemed to be the brother, son, wife, sister or father of another title, and its holders would treat each accordingly, regardless of their family relationship. Alliances were recorded in oral tradition as marriages, with the senior title described as the husband, the junior title, although held by a male chief, described as the wife.

---

So, it really does depend. It should be pointed out that in many cases, information from written sources first existed as oral information (see Herodotus or Bede) and was susceptible to the same issues described above. Being written down doesn't make a source inherently more trustworthy, but it does "freeze" the process of reinterpretation and recollection at the time it was written down.

Oral traditions do have shortcomings, and I talk a little more about that here. But, if you lack for written sources oral tradition can be an important source of information, though when possible it ought to be checked against linguistic or archaeological data.

I highly recommend you check out Jan Vansina's Oral Tradition as History. He provides a very thorough analysis of oral tradition, and its uses and shortcomings as a source for historical analysis.