The fairness debate


#1912

No, you will not see the less common occur more frequently. You will see the same frequency but in a more condensed timeframe. A 1:1000 event should still occur 1 in 1000 times but the 1000 events may take 50 hours live as opposed to 16.67 hours online. Frequencies of results are independent of how often the event being looked at occurs per X amount of time.

Interesting thought that the greater the number of instances per hour (despite the same frequency) could affect someone’s perceptions though. I hadn’t thought about that before and will need to let that idea germinate for a while to see if it makes sense. It sounds right from 1st read even though it doesn’t actually matter from a mathematical standpoint.


#1913

You are correct, I should have chose my wording a bit better. thank you :nerd_face:


#1914

I am a nerd about this stuff but words matter and definitions matter.

On the idea of repetition, do you have any sources you could point me to so I can read a bit more about it? The idea that people may get different impressions from identical frequencies occurring at different speeds sounds very interesting. I don’t know of any studies on how this has been observed and I’d like to read about it.


#1915

I guess someone who receives compensation for promoting a point of view cannot possibly be biased…but is free to assign various biasses to dozens of others because they hold a different point of view.

I’ll put this bluntly.
I do not accept the idea that it is appropriate for a Replay Rep to suggest that some kind of bias has colored my opinions, but it is inappropriate for me to suggest that a Rep’s opinion might be biased by the fact that they are receiving compensation from Replay.
If silencing people is how this debate is going to proceed, why have it at all?


#1916

I will leave you to do whatever you choose, You have your mind set. And the fact that I volunteered to be a Rep gives you more options to state your opinions, then so be it. I won’t fall for this. Just stating things that you can look up yourself. I’d start with college papers from Harvard, or Princeton under the topic of “Bias Studies”


#1918

May I correct one thing for you Warlock? :nerd_face:

I’m pretty sure the post where Juicee said “I won’t fall for this” and “your mind is made up” wasn’t in reply to you, but to:

I could be mistaken, but that’s how it sounded.


#1919

Then except friends request like I sent and take it private, out of context doesn’t work for me, “I won’t fall” isn’t directed at you


#1920

Didn’t see a friend request and I’m sorry if I mistook that comment as meant for me. It followed my asking a question about where to find these studies so I assumed it was. Post has been deleted by me.


#1921

It would be my honor if you would do it more frequently than that :slight_smile:


#1922

I can try. But it won’t be easy. You don’t make that many mistakes :joy:


#1923

It’s a burden but I’ve learned to live with it :wink:

Reminds me of an old joke: Once I thought I was wrong, but I was mistaken.


#1924

What I want is for my posts to not be deleted for suggesting bias when I was replying to a post that suggested bias, but wasn’t deleted.

Just be fair with the censorship.


#1925

Your post went to my email, and I thought it was funny and did not get offended, But I did see how some, possibly many would have. Not because you were attacking me, but because you were hitting a whole group. If you think “I am Bias” that would be fine (I’m not, and really don’t worry about what others think) but you can’t classify others as if they were me. hope that clears up things for you.


#1926

I ran some stats based on the hands that I can see on my activity page. Maybe this will be helpful.

Total hands 2230
Flops seen 2069
Turns seen 1752
Rivers seen 1445

Within this sample, here are some examples types of boards and how often each showed up. I included the expected # and % of each type.

Times seen % happened Times expected % expected
Flop with 2-flush 1115 53.89% 1139 55.06%
Monotone aka 3-flush flop 113 5.46% 107 5.18%
4-flush on board by river 68 4.71% 62 4.29%
5-flush on board by river 2 0.14% 3 0.21%
Paired flop 336 16.24% 351 16.94%
Flop all same value 2 0.10% 5 0.24%

#1927

Updated numbers. I added in the number of boards with 3 cards to a flush showing by the river. I also corrected my count of boards with a 4-flush - my total before incorrectly included turns with a 4-flush that didn’t reach the river.

Times seen % happened Times expected % expected
Flop with 2-flush 1115 53.89% 1139 55.06%
Monotone flop 113 5.46% 107 5.18%
River board with 3-flush 472 32.66% 496 34.33%
River board with 4-flush 63 4.36% 62 4.29%
River board with 5-flush 2 0.14% 3 0.21%
Paired flop 336 16.24% 351 16.94%
Flop all same value 2 0.10% 5 0.24%

Overall it doesn’t seem like there is anything here that is out of line from what you’d expect from random chance.


#1928

You are a data-geek! I knew I liked you.

How did you track this? Manually? Very small sample size to get a handle on the lower probability events but its informative nonetheless. The only large deviation from expected values came from the 3-of-a-kind flops - seen at less than half the expected rate. Because they are expected to occur so infrequently, there isn’t much significance in that finding.


#1929

Haha, I love numbers, it’s true. However, even I am not number-crazy enough to count thousands of flops by hand.

From the activity page, I expanded the sessions so that all of the hands were loaded. Then I copied the source HTML into Excel. Excel magic from there to manipulate it into something usable for calculating.

Agreed on the sample size. It would be much better to run this with say 100k hands instead.

I was expecting to see, for example, 3-flush boards occurring at a slightly higher rate than expected because people so love to chase flushes meaning boards with that potential are more likely to get to the river. On the other hand, there are removal effects - there are fewer possible flush cards to fall if there are flush cards in someone’s hand - so maybe this cancels out to some degree. All of my expected probabilities were calculated based simply on drawing 5 cards at random from the deck.


#1930

Have you written a little tool for excel? I would be very interested.


#1931

I’d like to know your magic too, ( with excell )


#1932

Here’s a Google spreadsheet version.

It’s read only - you can go File > Make a copy to get your own editable version. Or download to Excel if you prefer.

To get the source data for the hands into the spreadsheet:

  1. Go to your activity page and expand the individual sessions so that the hand history detail is shown. This only works for ring games because there’s no hand history detail available for tournaments.
  2. Then right click on one of the hands near the top and choose Inspect. I am using Chrome - the options might be different in other browsers.
  3. In the resulting screen, find the line that says ul class=“activity-games-list”. Right click it and go Copy > outerHTML.
  4. Now open the spreadsheet, select cell A2 on the first sheet, then paste.
  5. If you have a lot of hands you will need to fill down the formulas in the first two sheets of the spreadsheet to make it work. I have only set up this spreadsheet for about 150 hands in order to keep the file size small for downloads.