Unexpected match of term against larger but unrelated document token (containing the term)

This is about term “ant” being matched against title field and tags attribute
title: “… antenna …”
tags " “Antenna”
(in both fields “ant” is highlighted)

is this possible something related to missing index language?

I have even set up disableTypoToleranceOnWords to include “ant” but doesn’t matter

also in the ranking debug info (“ranking criteria for result x”) it states that no typo as been found/applied)

“ant” is matching as a partial word match in both cases, not as a typo. This is expected behavior for “search-as-you-type” experiences. What’s your desired outcome? Are there other records that include the word “ant” instead of “antenna” that should be ranked higher? Do you want to delay search until the fourth character? That would resolve the matching but degrade the end-user experience…

In my opinion there are 2 main search contexts:
(1) “search-as-you-type” or auto-complete and
(2) proper search (entering a term followed by “hitting enter”

in the auto-complete context I agree it makes sense to match “ant” and “antenna” but not all searches are auto-complete?

should we always wrap the last token in double quotes to prevent these matches from happening?

thanks for your response BTW

I think for this particular example, your vest bet is to modify the Exact criterion – “ant” isn’t matching as a typo, but rather as an inexact match since they are the starting letters in “antenna”.

You can also disable prefix matching to get the effect you describe without quotes, but we typically find customers prefer the “type-as-you-go” experience as it’s what they’re use to on larger sites like Google or Amazon.

Thanks for this. Some good tips in there.

I understand re auto-complete but our site doesn’t use it at all.

The 2nd link is the exact answer to my question