Skip to main content
Glama

Backtesting Arena

by io.github.Schoasch
OverviewDiscussionsAdmin

Historical analog — conditional forward returns

arena_get_historical_analog

What happened historically after the Bitcoin cycle looked like this? Conditional forward-return distribution for a named preset cycle state — over N DISTINCT historical episodes matching that state (matched_episodes), returns median/IQR/positive-share forward returns (30/90/180/365d) with per-horizon n, small-n warnings, point-in-time integrity and an evidence block that names which field its sample-size gate checked (gate_applies_to), against which threshold, over which data window. A distribution with its sample size. Not obtainable from web search or public market-data APIs — requires point-in-time indicator history and look-ahead-free episode matching. Presets: cycle_bottom_cluster (Cycle bottom cluster), cycle_top_cluster (Cycle top cluster), deep_fear (Deep fear), euphoria (Euphoria), quiet_volatility (Quiet volatility regime). The response opens with "preset_definition" (machine-readable condition set) plus current_state_matches (does the state hold TODAY?) and last_matching_date. Some presets carry a "study_finding" field — a state already investigated, with a NULL result where that is what the study found. EVERY preset returns "vs_unconditional_drift": the raw forward median contains the asset's contemporaneous drift; the drift and excess columns separate the two, and the excess can be negative while the raw median is positive. For quiet_volatility, vol_rank_threshold (fixed steps 5/10/20/50) asks the stricter "UNUSUALLY quiet" question the null study left open, and condition_on_direction conditions episodes on the sign of the first post-anchor move over direction_window_days (default 5) — both mark study_finding_applies=false, and horizons within direction_window_days are suppressed as circular. Also works for asset=ETH/SOL (F2 cycle history), but only price-derived presets (cycle_bottom_cluster, cycle_top_cluster) — fear-greed and volatility presets are BTC-only. Related: arena_get_volatility_history (the series behind the volatility preset), arena_get_cycle (the current state to compare against), arena_dip_scenario (composes this base rate into a tranche structure). [API Pro tier]

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
assetNoWhich asset’s cycle history. Default BTC. ETH/SOL support only price-derived presets (cycle_bottom_cluster, cycle_top_cluster).
presetYesNamed ex-ante cycle-state condition set. One of: cycle_bottom_cluster, cycle_top_cluster, deep_fear, euphoria, quiet_volatility.
forward_horizonsNoForward-return horizons in days. Default [30, 90, 180, 365] — except for quiet_volatility, which defaults to the horizons its study actually tested ([30, 90, 180]); anything beyond that is flagged as outside the protocol.
vol_rank_thresholdNoquiet_volatility only. Reference threshold as a FIXED step: 50 (default, below trailing median — the studied definition) or 5/10/20 (unusually quiet: RV30 below its trailing Nth percentile). Any value other than 50 sets study_finding_applies=false — the null study covered only the default.
direction_window_daysNoClassification window for condition_on_direction (default 5). Only meaningful together with condition_on_direction.
condition_on_directionNoquiet_volatility only. Condition episodes on the direction of the FIRST post-anchor move (sign of the direction_window_days-day return). Horizons <= direction_window_days are suppressed as circular. Sets study_finding_applies=false.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • addedInput schema / additionalProperties
      Added value: +false
    • removedInput schema / properties / context
      Removed value: -{
-  "description": "Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): \"Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization.\"",
-  "type": "string"
-}
    • changedInput schema / required
      Previous value: -[
-  "preset",
-  "context"
-]New value: +[
+  "preset"
+]
  2. Changed3 schema fields changed
    • addedInput schema / properties / condition_on_direction
      Added value: +{
+  "description": "quiet_volatility only. Condition episodes on the direction of the FIRST post-anchor move (sign of the direction_window_days-day return). Horizons <= direction_window_days are suppressed as circular. Sets study_finding_applies=false.",
+  "enum": [
+    "up",
+    "down"
+  ],
+  "type": "string"
+}
    • addedInput schema / properties / direction_window_days
      Added value: +{
+  "description": "Classification window for condition_on_direction (default 5). Only meaningful together with condition_on_direction.",
+  "maximum": 90,
+  "minimum": 1,
+  "type": "integer"
+}
    • addedInput schema / properties / vol_rank_threshold
      Added value: +{
+  "description": "quiet_volatility only. Reference threshold as a FIXED step: 50 (default, below trailing median — the studied definition) or 5/10/20 (unusually quiet: RV30 below its trailing Nth percentile). Any value other than 50 sets study_finding_applies=false — the null study covered only the default.",
+  "enum": [
+    5,
+    10,
+    20,
+    50
+  ],
+  "type": "number"
+}
  3. Changed3 schema fields changed
    • changedInput schema / properties / forward_horizons / description
      Previous value: -"Forward-return horizons in days. Default [30, 90, 180, 365]."New value: +"Forward-return horizons in days. Default [30, 90, 180, 365] — except for quiet_volatility, which defaults to the horizons its study actually tested ([30, 90, 180]); anything beyond that is flagged as outside the protocol."
    • changedInput schema / properties / preset / description
      Previous value: -"Named ex-ante cycle-state condition set. One of: cycle_bottom_cluster, cycle_top_cluster, deep_fear, euphoria."New value: +"Named ex-ante cycle-state condition set. One of: cycle_bottom_cluster, cycle_top_cluster, deep_fear, euphoria, quiet_volatility."
    • changedInput schema / properties / preset / enum
      Previous value: -[
-  "cycle_bottom_cluster",
-  "cycle_top_cluster",
-  "deep_fear",
-  "euphoria"
-]New value: +[
+  "cycle_bottom_cluster",
+  "cycle_top_cluster",
+  "deep_fear",
+  "euphoria",
+  "quiet_volatility"
+]
  4. First observed

TDQS

A4.9/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries full disclosure responsibility, and it delivers. It transparently discloses point-in-time integrity, look-ahead-free episode matching, small-n warnings, the 'study_finding' field with NULL results, the vs_unconditional_drift separation and its interpretation (excess can be negative while raw median is positive), circularity suppression for horizons within direction_window_days, and the fact that study_finding_applies=false when non-default parameters are used. It also clarifies that quiet_volatility's vol_rank_threshold has fixed steps and different semantics. No contradictions with any structured metadata exist.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is long but exceptionally information-dense and front-loaded. It opens with the core purpose, then systematically covers output fields, parameter nuances, asset limitations, and related tools. While it could be broken into bullet points for scannability, every sentence carries unique information and none is filler. The length is justified by the tool's complexity (6 parameters, 5 presets, multiple conditional behaviors). It loses a point only for being a single dense paragraph that might overwhelm an agent scanning quickly.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool has no output schema and no annotations, the description must compensate for both. It does so comprehensively: it explains the return structure (median/IQR, positive-share, per-horizon n, small-n warnings, evidence block with gate_applies_to), the preset definitions and their semantics, asset support restrictions, the special quiet_volatility parameters and their implications, and the relationship to unconditional drift. It even notes which presets carry study_finding fields. For an agent to call this tool correctly and interpret results, nothing critical is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Even though the input schema covers 100% of parameters with descriptions, the tool description adds substantial meaning beyond the schema. It explains the conceptual meaning of presets (e.g., 'cycle_bottom_cluster' as a named condition set), the special default behavior for forward_horizons under quiet_volatility, the semantic difference between vol_rank_threshold values (50 vs 5/10/20) and the 'unusually quiet' interpretation, and the directional conditioning semantics. The description enriches the schema with protocol-level nuance (e.g., horizons beyond study are flagged as outside protocol) that is not present in the schema fields.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a concrete question ('What happened historically after the Bitcoin cycle looked like this?') and defines the tool as a 'Conditional forward-return distribution for a named preset cycle state.' It specifies the verb (returns a distribution), the resource (historical analog episodes), and the distinguishing feature (point-in-time, look-ahead-free matching). It also names three sibling tools and when to use them, clearly differentiating itself from the broader family of arena_get_* dataset tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description states explicit usage context: it is not obtainable from web search or public APIs and requires point-in-time history and look-ahead-free episode matching. It names related tools (arena_get_volatility_history, arena_get_cycle, arena_dip_scenario) and states when each is appropriate. It also spells out asset restrictions (ETH/SOL only for price-derived presets) and parameter-specific usage (vol_rank_threshold, condition_on_direction only for quiet_volatility). No ambiguity remains about when to invoke this tool versus alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Backtesting Arena toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources