Algorithm

Suggestions & Feedback
41

A Nordic Warrior Champion knows what is trending on Replay.

I’d rather you strive to become a Nordic Warrior Champion than attempting

to become a Replay Billionaire.

Think about it.

How many Replay billionaires are there compared to Nordic Warrior Champions.?

Those billionaires are a dime a dozen.

Is that what you want to become?

I’m trending ahead of you in tournaments of late.

That’s a very nice trend to pick up on!

Of course, you’ll deny that trend exists too.

Say it is "an infinitesimally small sample of the massive, constantly growing

dataset that is Younguru."

All you’ve said, has already been stated in past algorithm discussions.

lihiue’s past words have more weight on this topic.

There’s a reason lihiue has more bank than you, he’s a better mathematician.

If you just tried a little harder, you “could’ve been a contender” and possibly become a

Nordic Warrior Champion!

It’s good to be a Nordic Warrior Champion! (Mel Brooks) LoL

Players may have 1B, 2B, 10B, 15B, but they don’t have a Nordic Warriors Championship.

The results speak for themselves.

Your narrow thinking about trends is very apparent.

Try a broader approach on multiple levels to get a grasp on what actually is

occurring right in front of your eyes!

Luke, would you agree a player applying trends will outperform a player who relies

solely on math?

I don’t think you will agree.

Yet, I’m trending ahead of you!

I’ve outplayed you time and time again.

If you just look at the recent results, right there in front of your eyes!

Maybe you should use your own intuitive pattern recognition skills to pick up on it.

image
image720×405 53.2 KB

As you know, the above was all in good fun.

I have great respect for the Younguru.

There can be a case made for Luke as the best ever on Replay Poker.

Rarely, if ever, did we see a steady and methodical climb to 1B

Something that Tiggy and Goatsoup never accomplished.

They were all over the map trying to get there.

Luke having that incredibly rare ability, obviously, the sky is the limit.

When you stop and think about it, 100M a month is the trend of all trends on Replay!

Yes, a Nordic Warrior Champion knows what is trending on Replay!

To put legendary in front of Younguru’s name, here’s an idea.

After you reach your goal of 1B, call your shot.

Declare you’ll win the first quarter 2025 of the Nordic Warrior League.

You’ll instantly become legendary, if you aren’t already legendary by now.

Then, call your shot again the next quarter.

1 Like
43

Replay’s most probable algorithmic weakness - lack of updates?

We’ve all seen how it plays out at the tables.

Entering first in a tournament certainly has a bias with more winning hands.

Twice I entered 1st recently and had a 2 ND and a 6 TH

Last night, the player entering 1 ST finished 1 ST in a crushing way.

It just happens way too often, even years back it was noticeable.

The best visual explanation I can give is it’s like a ride on the Tilt A Whirl.

One seat on the ride can whiz around like crazy, while the other seats stagnate

throughout the ride.

One spinning connection is greased up a little more.

Possibly two at each poker table but entering first seems to have the upper hand.

1. Seed Value Vulnerabilities

  • Seed Predictability: If the seed value used by the PRNG is not sufficiently random or is predictable, it can be exploited. For example, if the seed is based on the system clock, a knowledgeable attacker might be able to guess the seed and predict the sequence of cards.
  • Insufficient Entropy: If the RNG does not use enough entropy (randomness) in generating the seed, the resulting sequences might show patterns that can be exploited.

2. Algorithmic Weaknesses

  • Poor Implementation: If the RNG algorithm is poorly implemented, it might produce patterns or biases that skilled players can detect and exploit.
  • Lack of Updates: Over time, if the RNG algorithm is not updated or improved, it might become vulnerable to new techniques or attacks that can predict its output.

3. Data Mining and Analysis

  • Historical Data: Players who collect and analyze large amounts of historical hand data might identify patterns or biases in the card distribution. This can be used to make more informed decisions during play.
1 Like
47

Those are all theoretical problems, but none of them would explain any of the issues people think are down to the RNG. They wouldn’t make it possible for a human to be able to predict anything, but would in theory enable a computer to be able to perfectly predict every shuffle.
In practice, it’s basically impossible to exploit any of those when it comes to poker. There’s been a lot of online poker cheating scandals, none of them have ever involved exploiting the RNG.

1 Like
48

I agree but there have been instances where RNGs (Random Number Generators)

were compromised.

  1. Dual_EC_DRBG: This was a cryptographic RNG standardized by NIST, which was later found to have a potential backdoor inserted by the NSA. This backdoor could allow someone with knowledge of the backdoor to predict the RNG’s output1.
49

Right, it’s just that with poker, you’re never seeing the entire RNG sequence (most of the deck is never visible). Also, you either have one RNG per table, which means you’re not getting much data, or an RNG across multiple tables, which means you need to track multiple tables, but also introduces some genuine randomness.
You would be able to figure out the internal state of any pRNG that’s not being reseeded eventually, but eventually is likely to be a very, very long time.
It’s obviously a bit easier if someone has deliberately inserted a backdoor though :slight_smile:

50

I hear poorly designed PRNGs can exhibit patterns or correlations that reduce the

quality of randomness.

Replay Passed Marsaglia’s “Diehard” tests. It is a strong indicator of the quality of their random number generator (RNG), but it may not be sufficient for all applications. Here are some key points to consider:

Strengths of Diehard Tests

  1. Comprehensive:
  1. Historical Significance:

Limitations

  1. Not Exhaustive:
  1. Evolving Standards:
  1. Application-Specific Requirements:

Passing the Diehard tests is a good starting point and indicates a high level of randomness. However, for critical applications, especially in cryptography, it’s advisable to use additional test suites and ensure the RNG meets all relevant standards and requirements.

51

Yes, but a human would never be able to detect these issues. It’s a problem when you’re talking about cryptography, but if you just want a game of poker where everyone is on an equal footing, then even the worst RNG’s are going to suffice.

1 Like
52

image
1924 - 2011

George Marsaglia established the lattice structure of linear congruential generators in the paper “Random numbers fall mainly in the planes”,[2] later termed Marsaglia’s theorem.[3] This phenomenon means that n-tuples with coordinates obtained from consecutive use of the generator will lie on a small number of equally spaced hyperplanes in n-dimensional space.[4] He also developed the diehard tests, a series of tests to determine whether or not a sequence of numbers have the statistical properties that could be expected from a random sequence. In 1995 he published a CD-ROM of random numbers, which included the diehard tests.[5]

His diehard paper came with the quotation “Nothing is random, only uncertain” attributed to Gail Gasram, though this name is simply the reverse of Marsaglia G, and so likely to be a pseudonym.

He also developed some of the most commonly used methods for generating random numbers and using them to produce random samples from various distributions. Some of the most widely used being the multiply-with-carry, subtract-with-borrow, xorshift, KISS and Mother methods for random numbers, and the ziggurat algorithm for generating normally or other unimodally distributed random variables.

Always interesting lihiue.

53

STOP, please")