Higher scores for results with extraneous words

Brian00 wrote a day ago: 1

If I run an API query: https://api.tvmaze.com/search/shows?q=English the highest scoring match is:

{
"score": 0.6992032,
"show": {
"id": 72758,
"url": "https://www.tvmaze.com/shows/72758/english-teacher",
"name": "English Teacher",

and the next highest scoring match is:

{
"score": 0.6981286,
"show": {
"id": 55215,
"url": "https://www.tvmaze.com/shows/55215/the-english",
"name": "The English",

So if we drop the article The from the results we have an exact match on what is the second highest scoring match. I would think that that second highest scoring match should actually be the highest scoring match since it matches 100% after dropping the article.

Thoughts?

TomSouthwell wrote a day ago: 1

I know search is something @david is in charge of. You will have to be patient for a reply, but he will get to you soon :)

david wrote 15 hours ago: 1

The search results are not just sorted by how well the query matches, but also by how popular the show is. Because there is no exact match without the "the", in this case the more popular show is allowed to come on top.

At some point in the past we actually considered shows an exact match even if the "the" was dropped, but that caused problems elsewhere, because sometimes there's a show both with and without the "the" (e.g. "the voice" and "voice").