What a high score here does not mean
Every ranked list invites one dangerous reading: that the project at the top is the one to buy. That is not what this score says, and putting a number on something makes the misreading more likely rather than less, so it is worth being blunt about it on its own page.
A high score means one thing. On 4 August 2026, the paperwork for that project was in good order. Nothing more than that, and specifically not that it is a sound purchase.
What the score actually measures
Five checks, all of them counting documents rather than buildings.
- Register Standing Score. A mark out of 100 that this site works out for every project from five things the government already publishes: whether the record names Dholera, how many quarterly reports were filed, whether the completion date declared at registration has passed, how many complaint entries are on the public ledger, and how many other projects the promoter has registered. It measures paperwork. It says nothing about build quality or about price.
Of the 33 registered projects, 11 score 90 or more, 12 have filed every quarterly report that has fallen due, and 13 carry at least one complaint entry on the public ledger. Those are the real numbers behind the ranking, and none of them is about construction.
The bias built into the score, stated plainly
The third of the five checks awards 20 points when the completion date declared at registration has not yet arrived, and zero when it has. That is a real signal, because a builder who blows past a date they gave the state has given you useful information. It also quietly rewards being new.
On this edition, 17 of the 33 declared completion dates have already passed. A project registered in the last two years usually has 20 points that an older project has lost, whatever either has actually built. Every project in the top band registered in 2022 or 2023 or 2024 or 2025. Read the top of the list with that in mind, and compare a 2021 registration against other 2021 registrations rather than against last year's.
We could have removed this criterion and produced a flatter, less useful score. Instead it stays and the bias is printed here, because a scoring rule with a stated weakness is more use than one with a hidden strength.
Five things this data cannot see at all
These are the questions people actually ask before buying land. Not one of them is answerable from the register, and no amount of arranging the data will change that.
| What you probably want to know | Why this site cannot tell you |
|---|---|
| What the plot costs | The government publishes no price on the public project record. Not a rate per square metre, not a total, not a payment plan. Any Dholera page showing you a price list took it from a broker or made it up. |
| How much has actually been built | The portal does hold progress percentages block by block, but they are not in this edition of the data, so nothing on this site is scored on them. A project can file every report on time and have nothing standing on the ground. |
| Whether anything has been handed over | There is no feed of completion certificates behind this site. A completion date that has not yet passed is a promise, not a delivery. |
| Whether the title is clean | Ownership, encumbrances and land class sit in the revenue records, which are a different system entirely. Registration with the regulator says nothing about them. |
| Whether the builder is any good | There is no field for that, and there never will be on this site. We publish what the state publishes, and the state does not rate companies. |
| Nothing on this site is scored on any of the above, because there is no public field to score it from. | |
What you are usually buying in Dholera
One fact changes how the whole register should be read. 29 of the 33 registered projects are plotted developments, which means the thing being sold is a marked out piece of land, not a finished home. On a plotted scheme the developer is committing to the layout and the internal infrastructure, and the building is yours to do later.
That is a legitimate product and it is what most of this market is. It also means that the phrase possession date carries a different meaning here from the one it has in an apartment project, and that a completion date on the record is a date for the layout works, not for a house you can move into. The type breakdown shows the split in full.
Where this register itself is thin
Three limits worth carrying with you, all of them in the government data rather than in our arithmetic.
- GujRERA leaves the project land area blank on 10 of the 33 records. Those blanks are printed as blanks and never filled in from another field.
- 3 of the 33 records never write the word Dholera. They are included on a town planning scheme reference alone, stamped Inferred, and should be treated as unconfirmed. They are named here.
- 1 project has no quarterly report due yet, so it cannot be scored on the heaviest of the five checks. Its score is expressed against a smaller total, and that smaller total is printed beside the figure rather than hidden.
The complete list of ten declared gaps is on the method page, and it travels inside the downloadable data as well, so anyone who reuses the numbers inherits the warnings with them.
What to do with the ranking instead
Use it as a filter, not as an answer. A project that is registered, files on time, and has no complaint entry has cleared the low bar that a serious business clears. That tells you where to start looking, and it tells you nothing about where to stop.
After that, the questions are the ordinary ones, and they are answered by documents rather than by websites: the title chain, the land class, the layout approval, what the payment schedule actually commits the seller to, and what happens if the infrastructure arrives later than promised. This site cannot help with any of them. It can tell you that the project exists on the government record, which is where every one of those questions has to start.
Back to the full list of 33How the score is worked outWhy a builder might be missing