DarwinBotsManual
Open the app EspañolES

Frequently asked questions

Short answers to the most common questions when using the app and writing bots: bots that don't move, populations that die out or don't evolve, where your data lives and what differs from the original.

About the simulation

Why doesn't my bot move?

It's almost always one of these three things, and we tested all three with the engine:

You wroteWhat happensIt should be
.up 10 storestore takes the value from below and the address from above: this writes something else somewhere else.10 .up store
10 .upp storeA misspelled sysvar is worth 0 and the command doesn't arrive. The editor warns you: “.upp is not a sysvar: the engine reads it as 0. Did you mean .up?”.10 .up store
cond 10 .up store stopWithout start, the command stays in the condition section and writes nothing.cond start 10 .up store stop

If the DNA looks right, inspect the bot: in Summary, Genes active this cycle shows whether the gene that moves is running, and in Memory you can check .up (see The inspector). A movement command is cleared every cycle, so you have to give it in every one (see The cycle and the order of actions). More typical mistakes are in Common mistakes.

Why does everyone die?

A bot dies when it runs out of energy (see Death and corpses). The most common causes:

In the built-in scenarios, except the league matches, costs are 0: there, nobody loses energy just for living, and almost every death is a bot that was hunted.

Why doesn't my bot reproduce?

Check that the gene that writes .repro is running (the energy condition is usually the culprit) and that the bot has some energy and body. The most common reason is that the spot where the child would be born is occupied: a sibling right next to it, a wall or the edge of the field. The command stays written and is retried every cycle. The full list is in Reproduction.

Why don't they evolve?

To speed things up, a bot can reproduce with .mrepro, which makes that child mutate ten times more. All of this is in Mutations, and a guided experiment is in Your first evolution experiment.

Why does the simulation slow down?

Because there are a lot of bots: each one runs its DNA and moves in every cycle. Without costs nobody starves, so a species that eats well grows without limit. On the right of the Observe bar you can see how many cycles per second you're getting. To hold the population back, charge costs or lower the Vegetable cap and Solar energy (see Experiment).

Can I repeat exactly the same simulation?

Yes: the same scenario with the same seed gives the same simulation. The seed is shown and changed in Experiment. A saved run that you resume carries on fine, but not identically to how it would have gone on without being saved. See Seeds and reproducibility.

About the app

Where is my data kept?

In the browser you use, on that computer. Your bots, runs, scenarios, tournaments and reports aren't uploaded to any server. That means:

To keep them safe or take them somewhere else, export them: Download .dbsim when you save a run, Export my bots and marks (.json) in Bots, Export in scenarios and in tournaments. It's all in Your data.

I had bots and tournaments in the classic interface. Do I lose them?

No. The first time you open the new app, it copies the bots and tournaments you had saved in the classic one and tells you what it imported; the classic one is left untouched. You can also repeat it from the library menu, with Import from the classic interface, and Compete accepts the tournament files exported from the classic one. See The classic interface.

Can I use a bot from the original DarwinBots?

Yes. The DNA language is the same: a bot written for 2.48.32 loads and runs its DNA as in the original. You have three ways:

  1. On Home, the From a file card: choose the .txt and it is seeded into the default world, with algae.
  2. In Observe, Seed › Paste the DNA….
  3. In Bots, + New bot, to save it in your library and edit it.

Before looking for it, check whether it's already in the Bestiary: the app ships 684 bots from the forum and the wiki. The only language rule that changed is the else after a start, which in the original never ran and in this version does (see Genes: cond, start, else and stop). What does change a little is the world, because of the fixed bugs (see the next question).

How does it differ from the original?

The engine is the same program, translated. We fixed 35 bugs in the original that could be fixed without changing the DNA language: a sense that was always worth 0, energy that was destroyed instead of passing to the body, virus shots that were charged twice, and others. That's why the numbers in a simulation don't exactly match those of the original. The complete list, with what was deliberately left as it was, is in Differences from the original 2.48.32.

In the interface, the new app is different from the Windows program, and some things from the original were left out (such as custom drawings for the bots or the background image). The classic interface is closer to the original: see The classic interface.

The app says “N DNA words are not recognized and are worth 0”.

When you seed a species, the app checks its DNA. If there are words that are neither numbers, nor sysvars, nor operators (almost always a misspelled sysvar), it tells you how many. The bot is seeded anyway, but those words are worth 0. Open it in the DNA editor to see which ones they are: it marks each one with its line and, when it can, suggests the right name. See Common mistakes.

Why does my bot shoot members of its own species?

Because the DNA doesn't know about species: it shoots whatever is in front of it. You have to add a signature comparison to the shot's condition, such as *.refeye *.myeye != (“the signature of what I see isn't mine”). See Common mistakes and A bot that recognizes its own species.

Where do I learn to write bots?

Start with The structure of a bot and follow the tutorials in order, beginning with A bot that moves. The DNA editor warns you about errors as you write, and with Test you see how your bot does without leaving its profile.