The Problem of Overfitting in Herpetoculture:

This is an extension of a carousel post I made to my social media not long ago. For the sake of clarity, the term "overfitting" is not my invention. It is a concept that exists within computer modeling, economics, and professional poker. I note this to make sure no one gets the wild idea that this is anything I've come up with; it's just a useful way of thinking that happens to overlap well into Herpetoculture.

Overfitting is what happens when someone builds a theory around a small number of experiences or observations and mistakes that for a broad general rule. While this is not at all unique to Herpetoculture, it is exceedingly common in our beloved niche.

For example: A keeper observes something in one enclosure, one clutch, or with one pair of animals and then builds a broad explanation around that event from that tiny sample size. The observation is real, but the conclusion often reaches way beyond the observation.

Common Examples of Overfitting:

1. A clutch of eggs molds, so mold is assumed to be the cause of failure. 

(In truth, oftentimes mold is the symptom, not the cause of failure)

2. An animal eats a particular food item, so it’s assumed that food must resemble something in the wild diet. 

(In reality, reptiles are adaptable and will respond to nutrient and protein-rich foods regardless of familiarity with anything similar)

3. A breeding pair behaves or produces a certain way, so the entire species is assumed to follow the same patterns.

(Trends within species are best viewed with as large a sample size as possible; small sample sizes are misleading) 

In each case, the observation and the assumption are reasonable. The error is treating a limited observation as enough to support a large conclusion. We can only fix problems we’ve correctly identified.

The Fix: More Data! Always More Data!

Continually seek examples and counterexamples. Your observations matter, but they’re one piece of the picture. Before settling on a conclusion, ask yourself:

“How might I be wrong about the conclusion I’m reaching?” 

“Could something other than what I think or see be causing this?”

“What do others more experienced than me have to say about my conclusion?”

Growth in Herpetoculture requires resisting the urge to build universal rules from limited samples and experiences. The most reliable and valuable interpretations are those that remain after every reasonable attempt to falsify them.