What Worked
7-0
Structural & serve edges
Non-form edge types were perfect
87.5%
Double-digit edge plays
7-1 when estimated edge > 10%
+8.20u
Days 8-12
7-0 late tournament when data was deepest
+4.38u
Medvedev thesis
Three evolving bets on one read
~+130
Avg winner odds
Finding mispriced lines, not favorites
What Failed
5-4
Pure form edges
Market already prices form efficiently
0-3
WTA dogs
Putintseva, Raducanu, Eala: -3u
6-2, 6-0
Eala disaster
1000 words of false confidence
+4900
Futures
Right read, two tiebreaks short of 49u
Core Approach
Elo model as sanity check, never primary driver
Market consensus to identify mispricing
H2H + surface + form over historical data
Devil's advocate every thesis before committing
WTA avoidance after Day 3
1u base sizing, scale for conviction