Pre-launch· Token not live yet — no rewards or treasury until launchWhat this means
← Model
Decision trees

TRAIN v0.04

Switches to gradient-boosted trees (depth 2, 120 trees), which can learn thresholds and interactions a linear model cannot.
Completedpreviously deployed

Definition

Model type
Decision-tree ensemble
Signals used
Market signals
Technical details
max_depth
2
subsample
0.8
n_estimators
120
learning_rate
0.05
min_samples_leaf
10
Random seed
1004
Config hash
302937219738b1b1
Created
05 Oct 04:24
Completed
05 Oct 04:24
Artifact
models/v0.04/model.json
sha256 93d3c88bdc9814d3…

Dataset & compute

Dataset
Solana Launch Dataset v3
694 snapshots · 139 launches · sha256 7e44dbd62410
Training runs
1
Compute
TRAIN training server (CPU)
GPU hours
—
no GPU used
CPU time
5.1 s
Provider cost
$0.00
nothing purchased

Benchmarks

held-out test set · change vs v0.03
Spotting crashes (−70% within 1h)
0.872
Spotting survivors (still traded after 1h)
0.770
Spotting 2× runs (within 1h)
0.794
Calling the 1h direction (down / flat / up)
52.2%
Spotting creator dumps (within 1h)
0.847
Spotting crashes (within 6h)
insufficient data
Spotting survivors (6h)
insufficient data
Calling the 6h direction
insufficient data
Spotting survivors (24h)
insufficient data
Spotting migrations (within 24h)
insufficient data

Validation metrics (during training)

collapse_1h
AUC 0.990
n=100
collapse_6h
—
n=0
reach_2x_1h
—
n=100
survival_1h
—
n=100
survival_6h
—
n=0
survival_24h
—
n=0
migration_24h
—
n=0
trajectory_1h
acc 99.0%
n=100
trajectory_6h
—
n=0
creator_exit_1h
AUC 0.620
n=39

Validation launches are separate from training launches and from the locked test set. Benchmarks above are the public numbers.

What this version relies on

held-out validation launches
This version was trained before TRAIN started measuring which signals each model relies on.

For each signal, how much the score drops when that signal is scrambled across launches. TRAIN learns these weights itself from the data; nobody hand-codes them.

Training runs

Attempt 1Succeeded
05 Oct 04:24 · 00:00:01 · 594 examples · $0.00

Timeline

  1. 05 Oct 06:05
    TRAIN v0.04 deployed — now live
    average score across 1 tests 0.500 vs 0.333 for the live v0.03
  2. 05 Oct 04:24
    TRAIN v0.04 not deployed; v0.01 stays live
    Reason: no shared benchmarks yet; keeping the live model.
  3. 05 Oct 04:24
    Benchmark completed: TRAIN v0.04
    No locked test set yet (not enough held-out launches with observed outcomes). Benchmarks will be computed when it is locked.
  4. 05 Oct 04:24
    Benchmark started: TRAIN v0.04
  5. 05 Oct 04:24
    Training completed: TRAIN v0.04
    00:00:01 · trained on 594 examples · $0 compute cost
  6. 05 Oct 04:24
    Training started: TRAIN v0.04
    Solana Launch Dataset v3 · decision-tree ensemble · TRAIN training server
  7. 05 Oct 04:24
    TRAIN v0.04 queued
    Switches to gradient-boosted trees (depth 2, 120 trees), which can learn thresholds and interactions a linear model cannot.