The seed is the difference between "trust us" and "check it"
Every online draw tool tells you it is random. None of them can answer the one objection that matters: how does anybody know the seed was not chosen after somebody had a look at the entries?
From inside the tool, they cannot. So this one lets you type the seed in.
Announce the source before entries close. The Lotto numbers from a stated Saturday. A number picked by whoever put the prize up, in front of witnesses. The closing FTSE figure on a given day. Anything public that did not exist yet when the list was finalised.
Then the draw cannot have been fixed to suit the outcome, and you are not asking anybody to take your word for it. That is worth more than any amount of assurance about the algorithm.
One prize per person, or tickets stay in?
Somebody bought five tickets. They win first prize. Are the other four still in the drum for second?
In a physical raffle, yes, obviously, the tickets are still in there. In an office collection, almost always no, because people mean one prize each. Both are perfectly reasonable and they give different winners.
Most tools pick one silently. This one asks, and writes the answer into the record, because the worst possible moment to discover which rule was used is after the same name has come out twice.
Reserves are drawn now, not later
Winners go quiet. They change jobs, they miss the email, they turn out to have entered twice with two spellings.
Redrawing weeks later, with a new seed, and no record of the original, is where a clean raffle becomes something people mutter about. Reserves drawn at the same time, in order, from the same seed, mean the fallback was already decided before anybody knew it would be needed.
The checksum, and what it honestly does
It is eight characters computed from the entry list, and it changes if a single entry is added, removed, reordered or has its ticket count edited.
Publishing it with the result shows that the list nobody else saw was the list that was used. It also lets you prove the list without publishing everybody's name, which matters if the entrants did not sign up to be listed publicly.
What it is not is a cryptographic hash. It catches editing and honest mistakes. It would not stop somebody determined enough to construct a different list producing the same eight characters. For a raffle at work that is the right amount of machinery, and if your draw needs more than that, the answer is the pre-announced seed rather than a longer checksum.
The drum is a real drum
Five tickets means five slots in the list, not a weighting applied behind the scenes.
It comes to the same numbers either way, and it matters for a different reason: anybody checking your draw expects to see a drum full of tickets, because that is what a raffle is. A tool that turns it into probability arithmetic behind the scenes is harder to verify and easier to disbelieve.
Common questions
Why can I type in my own seed?
Because it is the only thing that makes a draw checkable by anybody. A seed the website generated can always be accused of having been chosen after somebody looked at the entries, and there is no answer to that from inside the tool. Announce a source before entries close, the Lotto numbers on a stated Saturday or a number picked by whoever put the prize up, then type it in. It could not have been fixed to suit the result, because it did not exist yet.
If somebody bought five tickets and wins, are their other four still in?
That is the question most raffle tools quietly decide for you, and both answers are defensible. In a physical drum the other tickets are still in there and they can win again. In an office raffle people usually mean one prize each. There is a switch, the record says which one was used, and the thing to avoid is finding out afterwards.
What is the checksum for?
It changes if a single entry is added, removed or edited, including a changed ticket count. Publishing it with the result shows that the list nobody saw was the one used. It is honestly a checksum rather than a cryptographic hash: it catches editing and mistakes, not somebody determined to construct a matching list. For a raffle at work that is the right amount of machinery, and if you need more than that the answer is the pre-announced seed, not a longer checksum.
Why draw reserves?
Because winners cannot always be contacted, and redrawing three weeks later with no record is exactly where a clean raffle turns into an argument. Reserves are drawn at the same time, in order, from the same seed, so the fallback was fixed before anybody knew it would be needed.
How do I give somebody more chances?
Put x5 after their name and they go into the drum five times. The tool builds a literal drum with one slot per ticket rather than doing weighted arithmetic behind the scenes, because that is what a raffle is and it is what anybody checking the draw expects to see.
What should I publish afterwards?
The full record. It carries the entrant count, the ticket total, the checksum, the seed, where the seed came from, which prize rule was used, and the winners and reserves in order. Anybody can put the same list and seed back in and get the same result. That turns a claim into something checkable.
Is the entry list sent anywhere?
No. It stays in your browser, there is no account and nothing is stored. Which also means the record is yours to publish or not, and the checksum is the part that lets you prove the list without publishing everybody's name.