Variance guide
A downswing is a result. Its cause needs more evidence.
PokerArc compares recent results with ranges generated from your recorded sessions. This describes how unusual a stretch is relative to that history.
Free tracking · No account required to start · Export your records
How the comparison works
Variance Lab uses bootstrap resampling: repeatedly drawing from recorded history to build comparison stretches. It uses 2,000 resamples. A percentile locates the recent result within that modeled distribution; it is not the probability that you played badly.
For example, a result near the fifth percentile is low relative to most modeled stretches. That alone cannot identify bad luck, mistakes, tougher opponents or a change in game selection.
Sample size and comparability matter
The interface requires at least eight sessions to calculate. That is a product minimum, not a guarantee of statistical reliability. A small or unrepresentative history can produce unstable comparisons. Sessions may also be correlated; changes in stakes, format, duration, currency or opponents weaken comparisons.
Historical resampling cannot recover outcomes absent from your data or promise future returns. Start by checking record accuracy and comparable games, then review decisions separately.
Start with a trustworthy ledger
Record sessions or import existing CSV history. With only one session, review totals and hours rather than drawing a variance conclusion.
For background on resampling, see the NIST/SEMATECH statistical handbook. PokerArc's model is a historical comparison, not a diagnosis of skill or a guarantee against loss.
Track your game. Understand your progress.
Record a session without signing up. Keep your data and export anytime.