GROUND TRUTH - SIMULATED dataset (Dataset B)
============================================
*** THIS IS SIMULATED DATA. Not from any patient. Do not report as real. ***
Regenerate with: Rscript scripts/00_simulate_dataset.R

TRUE clock rate    : 2.20e-03 substitutions/site/year
TRUE tMRCA (root)  : 1992.70
Epidemic origin    : 1990.0
Sampling window    : 2005.04 - 2025.00
Tips               : 60
Alignment length   : 1500 bp
Substitution model : GTR, no rate heterogeneity, strict clock

PLANTED METADATA ERROR (Exercise 6 - outlier detection):
  Tip          : SIM_ZMB_014_2022.97
  RECORDED date: 2022.97  (what students are given)
  TRUE date    : 2007.97  (used to simulate the sequence)
  Error        : +15 years
  This tip should show a large NEGATIVE root-to-tip residual: it claims to
  be recent but carries the divergence of an old sample. Removing it should
  visibly improve R^2. It is the ONLY planted date error in Dataset B.

TRUE transmission clusters : 15

  Cluster_A (n=2):
    SIM_BWA_004_2013.49
    SIM_BWA_005_2013.85
  Cluster_B (n=6):
    SIM_ZAF_007_2020.31
    SIM_ZAF_008_2018.76
    SIM_ZAF_009_2017.83
    SIM_ZAF_010_2016.08
    SIM_ZAF_011_2016.97
    SIM_ZAF_012_2013.66
  Cluster_C (n=3):
    SIM_ZMB_015_2024.83
    SIM_ZMB_016_2025.00
    SIM_ZMB_017_2020.39
  Cluster_D (n=4):
    SIM_BWA_020_2010.15
    SIM_BWA_021_2010.18
    SIM_BWA_022_2012.21
    SIM_BWA_023_2013.00
  Cluster_E (n=4):
    SIM_ZAF_025_2011.69
    SIM_ZAF_026_2015.08
    SIM_ZAF_027_2017.34
    SIM_ZAF_028_2017.13
  Cluster_F (n=4):
    SIM_ZMB_030_2021.92
    SIM_ZMB_031_2022.23
    SIM_ZMB_032_2021.78
    SIM_ZMB_033_2023.02
  Cluster_G (n=2):
    SIM_BWA_034_2025.00
    SIM_BWA_035_2019.68
  Cluster_H (n=2):
    SIM_ZAF_037_2018.14
    SIM_ZAF_038_2020.08
  Cluster_I (n=2):
    SIM_ZMB_040_2007.69
    SIM_ZMB_041_2006.40
  Cluster_J (n=2):
    SIM_BWA_042_2006.69
    SIM_BWA_043_2009.21
  Cluster_K (n=2):
    SIM_ZAF_046_2010.14
    SIM_ZAF_047_2015.98
  Cluster_L (n=2):
    SIM_ZMB_049_2007.36
    SIM_ZMB_050_2007.37
  Cluster_M (n=2):
    SIM_BWA_051_2007.15
    SIM_BWA_052_2007.97
  Cluster_N (n=3):
    SIM_ZAF_053_2020.84
    SIM_ZAF_054_2021.47
    SIM_ZAF_055_2015.21
  Cluster_O (n=2):
    SIM_ZMB_057_2022.51
    SIM_ZMB_058_2021.80

EXPECTED STUDENT RESULTS (vary slightly - estimation noise is normal):
  Root-to-tip R^2 : ~0.7 - 0.95   (strong temporal signal)
  TreeTime rate   : should bracket 2.2e-3 (roughly 1.7e-3 - 2.7e-3)
  TreeTime tMRCA  : should bracket 1992.7 (roughly +/- 5 years)

Marking notes:
  - A rate 10x off usually means the student used Dataset A's dates.tsv.
  - tMRCA is estimated by EXTRAPOLATION back beyond the oldest sample,
    so it always has wider error than the rate. Reward a sensible CI,
    not a bullseye.
