Σύγκριση μεθόδων

	Δοκιμή A/B διασταυρούμενης σχεδίασης (Crossover A/B Test)	Adaptive A/B Test
Πεδίο	Πειραματικός Σχεδιασμός	Πειραματικός Σχεδιασμός
Οικογένεια	Process / pipeline	Process / pipeline
Έτος προέλευσης	1949 (crossover design); 2000s (online A/B application)	1952 (Robbins); applied to A/B testing from ~2010s onward
Δημιουργός	Crossover design: E. J. Williams (1949); A/B testing framework: Ronald Fisher (experimental roots); modern online application widely attributed to Google and Microsoft experimentation teams	Herbert Robbins (bandit framework); Thompson Sampling formalized by William R. Thompson
Τύπος	Within-subject controlled experiment	Adaptive experimental design
Θεμελιώδης πηγή	Jones, B., & Kenward, M. G. (2014). Design and Analysis of Cross-Over Trials (3rd ed.). Chapman and Hall/CRC. ISBN: 9781439861424	Russo, D., Van Roy, B., Kazerouni, A., Osband, I., & Wen, Z. (2018). A Tutorial on Thompson Sampling. Foundations and Trends in Machine Learning, 11(1), 1–96.
Εναλλακτικές ονομασίες	within-subject A/B test, crossover split test, repeated-measures A/B test, AB crossover experiment	adaptive AB test, bandit A/B test, multi-armed bandit testing, online adaptive experiment
Συναφείς	6	6
Σύνοψη	A crossover A/B test is an experimental design in which the same participants or units are exposed to both treatment A and treatment B in sequence, with each serving as their own control. By eliminating between-subject variability, the design achieves higher statistical power than a standard parallel A/B test at the same sample size, but it requires careful handling of carryover effects and time-period confounds.	An Adaptive A/B test is an experimental design that dynamically reallocates traffic or participants toward better-performing variants during the experiment itself, rather than holding allocations fixed until the end. Drawing on multi-armed bandit algorithms such as Thompson Sampling or Upper Confidence Bound (UCB), it balances the exploration of uncertain variants with the exploitation of those already showing superior performance, typically yielding higher aggregate outcomes while still producing valid inferential conclusions.