Compare the rules and historical replay your strategy needs

TL;DR

NexusTrade fits experiments needing minute observations, historical option contracts or previous filled-order state; Composer fits visual stock and ETF allocation trees. Both support AI-assisted iteration. Equal-looking percentages can still produce different positions: sequential cash buys differ from target weights. The October 13, 2026 Composer sign-in reached identity verification before the builder, so this page supplies capability evidence and NexusTrade receipts rather than matched platform returns.

Evaluate a cooldown and a price condition against the same fill

Configure both indicators for the same asset and Buy side, with orderStatus Filled. DaysSinceOrder measures elapsed seconds divided by 86,400 from the matching fill time. LastOrderPrice reads the matching filled price. A submitted order cannot stand in for that fill.

The first entry needs its own condition. With no order history the time indicator returns infinity; LastOrderPrice also returns infinity for an empty history and can return no value when a nonempty history has no matching order. Do not rely on the cooldown expression alone to describe an initial-entry policy.

Composer’s conditions and filters can express complex market rules. For this exact previous-fill requirement, request an equivalent rule and its evaluated output before treating two strategies as interchangeable.

Choose the simulation by the rule you need to test

NexusTrade fits an experiment that depends on minute-level observations, historical option contracts or previous order state. Composer fits visually authored stock/ETF allocation trees and automated symphony execution. Both support strategy iteration; inspect the actual inputs and order semantics before choosing.

Composer supports sophisticated nested conditions. Its March 2026 release added ANY/ALL combinations, and its February research discussion describes AI and walk-forward tooling built on its API. The specific distinction here is fill-state access and replay data, rather than a general claim about complexity or access to AI.

Sign-in reached identity verification before the strategy builder

Composer's currentofficial workshop description reports that its MCP server was retired on April 17, 2026. It directs developers to the Composer API and users to Trade with AI for building and testing symphonies inside the app. The video was uploaded November 26, 2025; this retirement notice is from its description checked October 13, 2026. The historical MCP demonstration is not evidence of current availability.

Composer'sbrokerage-application help page says an approved brokerage account is required before creating, discovering or backtesting symphonies. It was last updated June 27, 2025 and remains published. Our observed session reached the identity form before the builder; the older policy alone does not establish every current account path.

SoFi'scurrent Composer disclosure identifies SoFi Securities and Composer Securities as separate registered broker-dealers. It does not establish that an existing SoFi identity verification carries over to Composer.

Inspect existing historical evidence before requesting another run

Claude Code created the inactive stock strategy, set its cadence, submitted one accepted test and read its completed result. Native MCP created the option test; Claude Code subsequently read it. The stock test demonstrates an elapsed fill-state cooldown; the separate price-plus-cooldown worksheet above is a constructed rule example.

For a minute experiment, choose interval Minute and explicit dates. The selected range plus minute-indicator warmup has a 730-day data cap. Keep the saved portfolio, capital, fees, benchmark and dividend policy fixed when comparing interval or rule changes.

NexusTrade also retains public spot-crypto strategy backtests. Composer’s February 3, 2026 release states that support for crypto assets and crypto indicators ended January 31, 2026. Older help/API examples that still mention crypto should not be treated as current availability.

Start with $10,000 and inspect the two allocation instructions

The saved NexusTrade recipe spends 85% and then 15% of remaining buying power. Composer's specified-weight blocks assign proportions of the symphony. These inputs produce different positions even when the labels both say 85/15. The worksheet assumes sequential fills, no fees and no price movement; it is sizing arithmetic rather than an executed order.

1. Build the target allocation in Composer

   In a symphony, choose specified weights under the top-level Weight block. Add QQQ and VOO Asset blocks and set 85% and 15%. Select the trading frequency separately. Inspect the resulting target allocation before backtesting or investing.

2. Read the NexusTrade instruction

   Open the saved recipe below. Its second buy uses the cash remaining after earlier fills; inspect the allocation type on each rule. It does not continuously restore target portfolio weights.

Keep the research input, rule and returned evidence together

Start with a read-only inspection of the saved portfolio and the evidence behind its inputs. Retain the source date and availability cutoff. Save one changed rule as a separate candidate, then approve its historical run with explicit dates and fees. Read query_backtest_status before interpreting query_backtest_history.

When a fresh test needs detailed signals and fills, generate_events:true costs 5 times the normal research tokens and its hot-store traces expire after 3 days. Read those events while retained and keep the configuration, completed statistics and source evidence together. API access alone does not establish that two platforms use the same data or fill model.

Composer also supports iteration through its visual editor and API. Match allocation units, observation interval, execution timing, costs and benchmark before comparing curves. No paired Composer/NexusTrade performance test was run for this page.

Inspect the rule in either visual editor

Both products have visual strategy authoring. NexusTrade's native condition builder edits nested rule groups, and its rule view displays the condition tree. Claude Code is another way to author the saved portfolio. The worked Minute run demonstrates a different historical cadence from Composer's documented daily backtest; it does not measure which visual editor is easier to use.

Inspect the complete saved QQQ and VOO JSON

[
  {
    "name": "Buy 85% of the available cash in QQQ every trading day.",
    "userId": null,
    "active": true,
    "condition": {
      "lhs": {
        "compound": false,
        "targetAssets": [],
        "value": 1,
        "type": "Value",
        "name": "Constant 1"
      },
      "rhs": {
        "compound": false,
        "targetAssets": [],
        "value": 1,
        "type": "Value",
        "name": "Constant 1"
      },
      "name": "always",
      "comparison": "equal",
      "type": "Base",
      "description": "Left-Hand indicator (comparator) Right-Hand Indicator",
      "example": "If the Rate of Change of Apple's price is > the Value 0."
    },
    "action": {
      "type": "Buy",
      "targetAsset": {
        "name": "QQQ",
        "type": "Stock",
        "symbol": "QQQ"
      },
      "amount": {
        "type": "percent of buying power",
        "amount": 85
      }
    },
    "orderExecution": {
      "type": "Market"
    },
    "automaticOrderApproval": false
  },
  {
    "name": "Buy 15% of the available cash in VOO every trading day.",
    "userId": null,
    "active": true,
    "condition": {
      "lhs": {
        "compound": false,
        "targetAssets": [],
        "value": 1,
        "type": "Value",
        "name": "Constant 1"
      },
      "rhs": {
        "compound": false,
        "targetAssets": [],
        "value": 1,
        "type": "Value",
        "name": "Constant 1"
      },
      "name": "always",
      "comparison": "equal",
      "type": "Base",
      "description": "Left-Hand indicator (comparator) Right-Hand Indicator",
      "example": "If the Rate of Change of Apple's price is > the Value 0."
    },
    "action": {
      "type": "Buy",
      "targetAsset": {
        "name": "VOO",
        "type": "Stock",
        "symbol": "VOO"
      },
      "amount": {
        "type": "percent of buying power",
        "amount": 15
      }
    },
    "orderExecution": {
      "type": "Market"
    },
    "automaticOrderApproval": false
  }
]

Continue exploring