Find Your Best Session, and Check Whether It Is Real
Posted: Sun Oct 25, 2026 1:22 pm
Find Your Best Session, and Check Whether It Is Real
A trader sorts 100 trades by session. Asian: 20 trades, total minus 5R. London: 45 trades, total plus 16R. New York: 35 trades, total plus 3R.
Averages: Asian minus 0.25R, London plus 0.36R, New York plus 0.09R. Total 14R over 100 trades.
The headline: London carries the account. Cutting Asian would have added 5R and removed 20 trades.
Test the claim. The standard error for London is 1.5 divided by the square root of 45, about 0.22R. Its average of 0.36 is 1.6 standard errors above zero. For Asian, the standard error is 0.34R, and the average of minus 0.25 is 0.7 standard errors below zero, which is weak evidence. A reasonable reading: London is probably good, Asian is probably not good, and New York is unclear.
Look for reasons. London has the deepest liquidity for the pairs traded, and the trader's strategy suits its trends. Asian hours are quieter for those pairs, and spreads are wider relative to range. A structural reason supports the pattern, which makes it more credible than a coincidence.
Action: reduce or stop Asian trading, with a plan to revisit after 40 more trades if curiosity remains. Keep New York, but look at it by hour.
Beware of cutting too finely. Splitting 100 trades into 24 hourly buckets leaves four per bucket.
Exercise: group your trades by session and compute the standard error of each average.
A trader sorts 100 trades by session. Asian: 20 trades, total minus 5R. London: 45 trades, total plus 16R. New York: 35 trades, total plus 3R.
Averages: Asian minus 0.25R, London plus 0.36R, New York plus 0.09R. Total 14R over 100 trades.
The headline: London carries the account. Cutting Asian would have added 5R and removed 20 trades.
Test the claim. The standard error for London is 1.5 divided by the square root of 45, about 0.22R. Its average of 0.36 is 1.6 standard errors above zero. For Asian, the standard error is 0.34R, and the average of minus 0.25 is 0.7 standard errors below zero, which is weak evidence. A reasonable reading: London is probably good, Asian is probably not good, and New York is unclear.
Look for reasons. London has the deepest liquidity for the pairs traded, and the trader's strategy suits its trends. Asian hours are quieter for those pairs, and spreads are wider relative to range. A structural reason supports the pattern, which makes it more credible than a coincidence.
Action: reduce or stop Asian trading, with a plan to revisit after 40 more trades if curiosity remains. Keep New York, but look at it by hour.
Beware of cutting too finely. Splitting 100 trades into 24 hourly buckets leaves four per bucket.
Exercise: group your trades by session and compute the standard error of each average.