Makati
Why three burger places came off this site in one afternoon
Three of the thirteen venues on the previous version of this site had permanently closed, including the one ranked first, because nothing was re-checking them. Every venue is now verified against the Google Places API monthly and anything reported closed comes off the page.
By Tony Hilden · · 3 min read
On 27 July 2026 we rebuilt this site from scratch and re-checked every venue on it against the Google Places API. Three of the thirteen came back reported as permanently closed. One of them was sitting at number one, and its photo was the image the site used when anyone shared a link to it.
That is the whole problem with a list nobody re-checks. It does not degrade visibly. It keeps looking authoritative while quietly sending people to a shut door.
What came off
Three venues were removed because Google reports them permanently closed. Two more were dropped for a different reason: we could not resolve them to a real trading business at all. One of those matched a hotel rather than a burger place when queried, which tells you something about how loosely the original list had been assembled.
That left nine verified venues from an original thirteen. A shorter list, and a true one.
What replaced the ratings
The previous version of this site displayed scores like 9.9 and 9.2. Google does not use a ten-point scale. Those numbers were generated, not measured, and there was no way for a reader to tell.
Every rating on the site now comes from the Places API, and each card states which source it came from and the date it was pulled. If a number here is stale, you can see exactly how stale.
How the ranking actually works
A raw average is a poor way to order a list. A venue with 4.7 stars from fifty people looks better than one with 4.4 from six hundred and thirty-five, but the second result is much harder to produce and far less likely to be a fluke.
So position comes from a Bayesian average that pulls low-volume ratings toward the category mean:
score = (v / (v + 150)) × R + (150 / (v + 150)) × 4.3
Where R is the venue’s rating and v is its review count. At 150 reviews a venue’s own average carries half the weight. Below that, it is pulled toward 4.3, the category average for local venues here. The full method is written out on the how we choose page.
The useful consequence is that a new place with a handful of glowing reviews cannot leapfrog somewhere that has held a good average across hundreds of visits. It has to earn the position with volume as well as quality.
What we will not do
We do not publish AI-generated photographs of real restaurants’ food. Where we have no photo we can legitimately use, the card shows the venue name set in type instead. Every photo that does appear comes from Google Places and carries the name of the person who took it.
We do not sell positions on these lists. Some venues here are LocalEnhance clients, and
those are labelled on their own cards, but no commercial arrangement moves anything up. If
a placement is ever paid for, it carries a visible label and a rel="sponsored" attribute
on its outbound link, so both readers and search engines can see it.
The check now runs every month. When something closes, it comes off.