A recent study conducted by the Indian Institute of Technology (Banaras Hindu University), Varanasi, has provided a scientific affirmation of the reliability and statistical integrity of Random Number Generators (RNGs) employed in online card games. The research was led by Dr. Bhaskar Biswas, Professor and Head of the Department of Computer Science & Engineering at IIT-BHU, and his team. They deployed internationally accepted methodologies to assess the randomness and fairness of card shuffling algorithms using real gameplay data.
Table of Contents
The study used the ‘Dieharder’ statistical test suite, a globally recognized benchmark for evaluating RNG quality, to evaluate thousands of card distribution sequences drawn from historical gameplay logs. Researchers rigorously tested three months of non-Personally Identifiable Information (non-PII) gameplay data obtained from RummyCulture, one of India’s leading online skill-based gaming platforms and the flagship platform of Gameskraft. The analysis spanned both 53-card and 106-card game formats, examining the RNG output across multiple dimensions of randomness including uniformity, independence, and unpredictability.
IIT-BHU Study Confirms High Statistical Fairness in Random Number Generators Used in Online Card Games
Key findings from the study include:
- For 53-card games, RNGs passed 97.34% of all Dieharder tests and for 106-card games, RNGs passed 98.25% of the tests.
- P-value distributions, an indicator of randomness quality, were found to be largely uniform across both formats, with only slight clustering in edge ranges (0.0–0.1 and 0.9–1.0), which is consistent with theoretical expectations in truly random datasets.
“These results demonstrate a high level of statistical conformity and provide reassurance about the robustness of RNG implementations in digital card games,” said Dr. Bhaskar Biswas. “In games such as online rummy, the underlying randomness must be beyond reproach. Our analysis shows that, when properly implemented, RNGs can uphold the core principles of fairness and unpredictability.”
This marked a pioneering moment for RummyCulture, making it the first instance in the skill based Real Money Gaming (RMG) and the broader global gaming ecosystem where an independent academic institution reviewed actual game data.
The IIT-BHU team validated the RNGs using Dieharder’s extensive suite, which includes tests such as:
- Birthday Spacing Test (to detect duplicate values in subsets),
- Bitstream and Rank Tests (to verify bit-level randomness),
- Parking Lot and Minimum Distance Tests (to simulate spatial distribution),
- DNA and OPSO/OQSO Tests (for detecting overlapping sequence anomalies).
Each of these tests was applied repeatedly with large datasets, and the results consistently affirmed the absence of bias or predictability in the number sequences generated.
Importantly, the report stresses the regulatory relevance of these findings. RNGs in digital games are often subject to audits and certification under frameworks like the NIST standards and the British Remote Technical Standards (RTS). The IIT-BHU analysis confirms that the RNG systems studied meet the statistical expectations laid out by these international benchmarks.
This independent study reaffirms RummyCulture’s long-standing commitment to fair play, including prior RNG certifications from globally competent authorities like iTech Labs(ISO 17025 certified), whose tests have also affirmed unpredictability, non-repeatability, and uniform distribution, consistent with Marsaglia’s ‘diehard’ tests for statistical randomness. Reports like these are essential for upholding ethical standards, building user confidence, and ensuring the long-term sustainability of the expanding online gaming industry.
Read more: From Cricket to Hockey: India vs Pakistan Asia Cup Drama Takes Explosive New Turn
