Lucky Skill

 

RNG Evaluation Report

 

 


1.     Operator

Lucky Skill Ltd.                                                URL: http://www.luckyskill.com

 

2.      Test house

iTech Labs Australia                                        URL: http://www.itechlabs.com.au

Suite 24, 40 Montclaire Ave                              e-mail: info@itechlabs.com.au

Glen Waverley

VIC 3150, Australia

 

3.      Software Provider

Lucky Skill Ltd.                                               URL: http://www.luckyskill.com

Address:                            

Atho 3 B-C

Agioi Omologites

P.C. 1087

Nicosia, Cyprus

 

4.      System/Module tested

System: N/A                                                   URL: N/A

Module: Random Number Generator (RNG) using Mersenne Twister algorithm.

Date Completed: 23 January 2009                            

 

 

5.      Previous history of items under test

None.

 

6.      Evaluation performed

                                                   iTech Labs has conducted evaluation for the RNG implementation using Mersenne Twister algorithm
                                                   as below:


This RNG consisted of implementation of Mersenne Twister algorithm. Our evaluation of the RNG consisted of source code evaluation, Diehard tests on the raw numbers generated by the algorithm and Chi-square tests on the shuffled decks for card games and dice numbers for the dice games.

1.        Source code examination
The following source code evaluation was conducted:

           a)   Identification of RNG algorithm;

           b)   Security of internal state, seeding and re-seeding, thread safety;

           c)   Shuffling of cards;

           d)   Scaling for the ranges required for the non-card games

2.        Tests conducted

           a)   Marsaglia’s “Diehard” tests were applied to 80 million bits of raw 32 bit random numbers        generated by the Mersenne Twister algorithm. The following diehard tests were conducted on 2        sets of 80 million bits;

                 i)      BIRTHDAY SPACINGS

                 ii)     OVERLAPPING 5-PERMUTATIONS

                 iii)     BINARY RANK TEST for 31x31 matrices

                 iv)     BINARY RANK TEST for 32x32 matrices

                 v)      BINARY RANK TEST for 6x8 matrices

                 vi)     BITSTREAM TESTS ON 20-BIT Words

                 vii)    BITSTREAM TESTS OPSO, OQSO, DNA

                 viii)   COUNT-THE-1's IN A STREAM OF BYTES

                 ix)     COUNT-THE-1's IN SPECIFIC BYTES

                 x)      PARKING LOT TEST

                 xi)     MINIMUM DISTANCE TEST

                 xii)    THE 3DSPHERES TEST

                 xiii)   THE SQEEZE test

                 xiv)   OVERLAPPING SUMS TEST

                 xv)    RUNS TEST

                 xvi)   CRAPS TEST

           b)   The following Chi-squared tests were conducted:

                 i)     Shuffling tests for 8 decks of cards. These tests were conducted using samples ranging from               1,000 to 100,000 deals for a total of over 2.5 million deals

                 ii)    Scaling tests for the Dice games.

           The RNG tests were conducted for compliance to relevant Alderney Gambling Control Commission (AGCC), UK Gambling Commission, Isle of Man and Malta standards.

 

7.     Evaluation results

1.         Source code examination.

Lucky Skill Ltd. RNG implements Mersenne Twister (MT) algorithm.
We identified a small number of issues including some regarding scaling and shuffle. All these issues ware resolved in the updated code provided by Lucky Skill Ltd.

2.        Tests conducted

           a)   Marsaglia’s “Diehard” tests
       The results were satisfactory.

           b)   Chi-squared tests

                 i)     Shuffling tests
      The results were satisfactory.

                 ii)     Scaling tests for the specified ranges
      The results were satisfactory.

 

8.     Observations

1.         Thread safety: If multiple threads access a single instance, access should be synchronized externally. Or alternately, different threads must use different instances of the RNG.

 

9.     Certification

Date of Certification: 23 January 2009                     

Software provider: Lucky Skill Ltd                      

Total number of pages: 6                                       


iTech Labs certifies that the RNG (listed in Appendix-A) comply with Alderney Gambling Control Commission (AGCC), UK Gambling Commission, Isle of Man and Malta standards subject to the conditions in section 10 Conditions.

 

 

10.       Conditions of the Certification

1.         The source code provided to iTech Labs (as per Appendix-A) must be used for compilation of the RNG module.

2.         While using the RNG, the issue identified in section 8 Observations should be satisfied.

2.         Any change to the RNG source files listed in Appendix-A must be verified by iTech Labs.

 

11.       Conclusion

While it is not possible to test all possible scenarios in a laboratory environment, iTech Labs has conducted a level of testing appropriate for a submission of this type.

 

Accordingly, subject to the above comment, iTech Labs certifies that the items under test comply with the relevant Technical Standards, unless otherwise stated.

 

 

 

 

                                

 

Geoff Nicoll                                                              Kiren Sreekumar

Principal Consultant                                                 Principal Consultant

iTech Labs Australia                                                iTech Labs Australia

 

23 January, 2009                                                          23 January, 2009

 

 

 

Appendix-A


                                             Click here to view the Original iTech Labs report.