Conditions

A condition is a rule that is true or false. The bot checks it at a fixed place in the strategy, and goes on only when it is true. A strategy has conditions in these places:

Rows that must all holdCopy Anchor Link

In the workflow, a condition is a list of rows. All rows must hold. Each row has:

1. A value, for example token.buyers.
2. Its window, if the value has one, for example 60 seconds.
3. A comparison.
4. The number or word to compare with.

Select Add a condition to add the first row, and And also to add more. Select the "x" on a row to remove it. A condition with no rows is "Always": the check goes on every time.

The comparisons are:

The yes or no values are token.graduated, token.creator_sold and event.first_buy, and AI answers or webhook fields of the yes or no type. All other rule values are numbers.

In JSON, a value is {"f": "<name>"}, with "args" for its window. A comparison has the value first and the number second. Rows that must all hold are an and:

{
  "and": [
    { ">=": [{ "f": "token.buyers", "args": [60] }, 30] },
    { "not": { "f": "token.creator_sold" } }
  ]
}

WindowsCopy Anchor Link

Some values count what happened in a recent window, for example token.buyers(window_s): the distinct wallets that bought in the last window_s seconds. The window is in seconds, and it is at most 3600 (one hour). The engine keeps one hour of trades and prices for each token.

token.price_change(window_s) is the price now divided by the price window_s seconds ago, so 1 is flat and 1.5 is up 50%. Until the engine has watched a pool for that long, it has no value.

Missing valuesCopy Anchor Link

A value can be unknown, for example token.age_s for a token that launched before the engine started to watch it. A comparison with an unknown value is false. So a rule over missing data never buys and never sells. You do not need to check for missing values.

PricesCopy Anchor Link

token.price is a price in wei per the token's smallest unit. For a token with 18 decimals (every Pons token), that is ETH per whole token.

The workflow shows and takes prices in ETH per whole token when it knows the token's decimals. The JSON keeps the stored number. For a token with fewer decimals, the stored number is larger: ETH per whole token times 10 to the power of (18 minus its decimals). Read prices per whole token.

Where each value can be usedCopy Anchor Link

Values come in four groups. Not every group can be used everywhere:

The editor offers only the values that fit where you are. The engine refuses a value that is used where it cannot be known.

The tokenCopy Anchor Link

The trade or launch that started the checkCopy Anchor Link

The workflow does not offer event.trader. You can use it in JSON.

The position (sell rules only)Copy Anchor Link

The botCopy Anchor Link

Values from AI answersCopy Anchor Link

Each AI step declares the fields of its answer. A condition after the AI steps can read them as ai.<step>.<field>, for example ai.decide.action. You can use them in Then buy if, in Then sell if, and in Run this step only if of a later step.

A rule can read fields that are a number, yes or no, or one of a list of words. It cannot read text, images or token addresses. A word must be one of the field's own words: the engine refuses "buyy" for a field whose words are buy and skip.

{
  "and": [
    { "==": [{ "f": "ai.decide.action" }, "buy"] },
    { ">=": [{ "f": "ai.decide.confidence" }, 0.7] },
    { "!=": [{ "f": "ai.decide.risk" }, "high"] }
  ]
}

No answer, an answer in the wrong shape, or an answer that comes too late never buys or sells. Read AI steps.

Values from webhook fieldsCopy Anchor Link

With a webhook trigger, entry conditions can read the fields your service sends as hook.<field>. As with AI answers, a rule can read numbers, yes or no, and words, but not text or token addresses. Read webhooks.

What only JSON can expressCopy Anchor Link

The workflow rows cover most conditions. JSON can also express these. The workflow keeps such a condition exactly as it is, and you edit it in the JSON tab.

OrCopy Anchor Link

or holds when at least one of its parts holds. This condition holds when 50 wallets bought in the last minute, or when more than 1 ETH net went in:

{
  "or": [
    { ">=": [{ "f": "token.buyers", "args": [60] }, 50] },
    { ">": [{ "f": "token.net_eth", "args": [60] }, 1] }
  ]
}

Not around more than one valueCopy Anchor Link

not around a single yes or no value is a row (is no). not around a larger condition needs JSON. This condition holds unless the token both has graduated and is more than an hour old:

{
  "not": {
    "and": [
      { "f": "token.graduated" },
      { ">": [{ "f": "token.age_s" }, 3600] }
    ]
  }
}

ArithmeticCopy Anchor Link

+, -, * and / combine two values. This condition holds when the token had more than twice as many buys as sells in the last 5 minutes:

{ ">": [{ "f": "token.buys", "args": [300] }, { "*": [{ "f": "token.sells", "args": [300] }, 2] }] }

A division by zero gives an unknown value, so the comparison is false.

A value on the rightCopy Anchor Link

The workflow rows put the value first. A comparison with a value on the right, or with a value on both sides, needs JSON.

Limits of a conditionCopy Anchor Link

• An and or an or has 1 to 20 parts.
• A value takes at most 4 arguments.
• A word or text is at most 200 characters.
• There is no null in rules. A missing value already makes a comparison false.