Sets on the river @ party

[Ace-Korea]

I play SnG tournaments at party and I just got busted out with QQ by JJ (we went all-in) because a Jack came on the river. As soon as I saw him holding JJ, I KNEW I was beat b/c he would make a set on the river... and he did.

If my calculation is correct, a set is made on the river about 4.3% of the time. (2 out of 46 cards left = 4.3%)

But I'm seeing it WAY MORE than 4.3%. It could be due to short-term fluctuations, but it seems like it happens in every game I play. (It also happened in the last game I played.) I'm starting to think that party's algorithm might be messed up or something.

So I want to ask SnG players who play at party whether they also experience too many sets on the river or not.

Remember, 4.3% means about 4 or 5 occurrences out of 100 times.

[eastbay]

[ QUOTE ]

Remember, 4.3% means about 4 or 5 occurrences out of 100 times.

[/ QUOTE ]

Gee, ya think?

What is this, another "party is rigged" post?

It isn't. Get over it.

eastbay

[benfranklin]

[ QUOTE ]

If my calculation is correct, a set is made on the river about 4.3% of the time. (2 out of 46 cards left = 4.3%)

But I'm seeing it WAY MORE than 4.3%. It could be due to short-term fluctuations, but it seems like it happens in every game I play. (It also happened in the last game I played.) I'm starting to think that party's algorithm might be messed up or something.


[/ QUOTE ]

How much more according to your records?
Here is an old post by Paul Phillips on rgp:

Some of you know that in a past life I was a programmer and a manager
thereof. In mid-1997 our company acquired an online game site called
playsite that had a decent population of people playing classic games,
one of which was backgammon. The codebase was something of a mess though,
so we undertook a complete rewrite and released it in early 1998.

After we released the new code, we began receiving email from people
and hearing chat online that there were unusually many doubles being
rolled in the backgammon games. That sounded unlikely but I took a
look into the code, and it was as straightforward as could be, no room
for a wacky error. The server picked two random numbers from 1 to 6
in the normal java fashion.

The java random call is a simple wrapper around the C library function.
We were seeding it in the normal ways. Everything was fine. But the
complaints were unrelenting, so we took increasingly extreme measures
trying to figure out what was going on. First we incorporated a java
RNG to avoid the C library. When this didn't "help", we started
logging all the die throws and did statistical analysis on tens of
thousands of logged rolls.

What we found was that doubles were being rolled at precisely the rate
one would expect. There was absolutely nothing surprising in the stats.
We communicated this to the complaining players, but it still didn't do
any good. You could go into a backgammon lobby anytime and you'd
rarely have to wait more than a couple minutes before chat would emerge
that "everyone knew" that too many doubles were being rolled. It had
entered the realm of known facts, and there was no getting around it.

We closed the dozens of filed bug reports involving our loaded dice and
moved on with our lives, but I've never forgotten the certainty with which
people asserted that our dice were not rolling right. And the point, of
course, is how similarly that certainty is echoed here when people talk
about online poker being rigged for this or that result.

I see three major factors contributing to this misplaced certainty.
The three are the same whether we're looking at original vs. rewritten
playsite, or B&M poker vs. online poker. Much of this has been
written before by myself and others, but I include it here to help
illustrate how similar the backgammon and online poker situations are.

1) SPEED. We build an unconscious model of how often noticeable events
take place, but it's largely rooted in time, not in number of events.
When the number of events per unit time increases (the rewritten playsite
was of course faster, just as online poker is faster than B&M) then we
are surprised to observe more noticeable events.

2) SELECTION BIAS. We notice quads. We notice doubles. We feel like
we know how often they happen because we know that we notice them, but we
do not know how often unnoticeable events take place. We therefore lack
the necessary data to do analysis, but we have so much faith in our brains
as pattern recognition machines, we try it anyway.

3) MEMETICS. This is in some ways the biggest one. When you're
surrounded by people who have become convinced that something is true,
it's difficult not to start believing it's true yourself. Online chat
environments make it very easy for people to share their feelings about
the injustice of the randomness, and it's such a seductive idea anyway,
it's not hard for it to gain followers. Read "The Tipping Point" for more.

In closing, here is one quote I found in my old email. I wish I had
the whole file so you could see how widespread the certainty was.

Message: your dice are throwing doubles again---CALIBRATE THEM! get
your act together

A message to online poker sites: Your decks are dealing bad beats
again. CALIBRATE THEM!

[Ace-Korea]

[ QUOTE ]


Gee, ya think?

What is this, another "party is rigged" post?

It isn't. Get over it.



[/ QUOTE ]

Actually, party doesn't generate "random" outcomes (hence the pattern map?). So in a sense it is rigged, IMO.

Oh, by the way, the very next game I played right after posting the message, I made sets on the river twice in the same game. 1st place, yay~ /images/graemlins/wink.gif


The Luckiest Guy in the World

[eastbay]

[ QUOTE ]
[ QUOTE ]


Gee, ya think?

What is this, another "party is rigged" post?

It isn't. Get over it.



[/ QUOTE ]

Actually, party doesn't generate "random" outcomes (hence the pattern map?). So in a sense it is rigged, IMO.


[/ QUOTE ]

Show me statistical evidence that any particular kind of hand happens more than its fair share on PP.

And "dude I swear it happens all the time" does not qualify.

eastbay

[AleoMagus]

[ QUOTE ]
Actually, party doesn't generate "random" outcomes (hence the pattern map?).

[/ QUOTE ]

HILARIOUS!
/images/graemlins/grin.gif

[Ace-Korea]

[ QUOTE ]


Show me statistical evidence that any particular kind of hand happens more than its fair share on PP.



[/ QUOTE ]

I'm thinking about it. Even with PokerTracker it wouldn't be an easy thing to do. But if I do it, it will be something like this.

1) Gather all hands where 2 or more players went all-in pre-flop and at least one of the players had a pocket pair.

2) Figure out how many times someone made a set on the river out of the total.

3) If it's relatively larger than 4.3% then I might be onto something, ASSUMING that my data is sufficiently big enough and 4.3% is correct.


I'll post the results by midnight.

[benfranklin]

[ QUOTE ]


And "dude I swear it happens all the time" does not qualify.

eastbay


[/ QUOTE ]

Dude, if he swears it happens, it must be like true. What are you, like some kind of like fact fanatic?

[PrayingMantis]

[ QUOTE ]
I play SnG tournaments at party and I just got busted out with QQ by JJ (we went all-in) because a Jack came on the river. As soon as I saw him holding JJ, I KNEW I was beat b/c he would make a set on the river... and he did.

If my calculation is correct, a set is made on the river about 4.3% of the time. (2 out of 46 cards left = 4.3%)


[/ QUOTE ]

I don't see any problem. It should happen 4.3% of the time, and it happened to you THIS time. So? Oh, I see, you KNEW it's going to happen. Well, you're actually suggesting it happens 100% of the time, or that you are a psychic and can see into the future. If it happens on Party, say, 4.7% of the time, instead of 4.3% (which would mean that their generator is COMPLETELY biased and nobody should play there, of course), would you still KNOW it's going to happen? Do you see my point?

Edit:

BTW - if you both have pocket-pairs, one of you will make a set on the river about 4.5454% of the time, if he didn't make it until then. You can subtract 2 more cards from the deck, to make your "outs" calculation.

[PrayingMantis]

[ QUOTE ]
BTW - if you both have pocket-pairs, one of you will make a set on the river about 4.5454% of the time, if he didn't make it until then. You can subtract 2 more cards from the deck, to make your "outs" calculation.

[/ QUOTE ]

To put it more accurately, *one* of you will make a set on the river 9.09% of the time (i.e., HE will make it 4.5454%, YOU will make it 4.5454%).

Yep, amazingly rare.

[Prickly Pete]

[ QUOTE ]
I'll post the results by midnight.

[/ QUOTE ]

I've withdrawn all my funds from Party while awaiting the results. Anybody know where I can find that properly calibrated Backgammon site?

[benfranklin]

[ QUOTE ]
[ QUOTE ]
I'll post the results by midnight.

[/ QUOTE ]

I've withdrawn all my funds from Party while awaiting the results. Anybody know where I can find that properly calibrated Backgammon site?


[/ QUOTE ]

No results by midnight. Guess he turned into a pumkin. How'd you do at Backgammon??

[Ace-Korea]

[ QUOTE ]
No results by midnight. Guess he turned into a pumkin.

[/ QUOTE ]

Sorry for being late. It's taking longer than I thought. (I'm doing it manually, one hand history file by one.)

So far, I'm getting about 9-11%. And I never said it happens ALL THE TIME. I said it happens far more often than 4.3%. Don't change the words.

I'd post the results as soon as I'm done with it.

[PrayingMantis]

[ QUOTE ]
So far, I'm getting about 9-11%

[/ QUOTE ]

Read my two posts in this thread, before you're wasting any more time on this holy mission of yours.

[Ace-Korea]

[ QUOTE ]
Read my two posts in this thread, before you're wasting any more time on this holy mission of yours.

[/ QUOTE ]

I know exactly what you're talking about but my calculation is still correct, IMO. You gotta trust me on this. I'm getting around 10% when it should be around 4%. I have taken everything into account. Just wait for more results cuz it's not significant yet. THANK YOU.

[TylerD]

Dude, your post is ridiculous, everyone knows its PokerStars that rigs the river not Party.

[benfranklin]

[ QUOTE ]
Dude, your post is ridiculous, everyone knows its PokerStars that rigs the river not Party.

[/ QUOTE ]

I always get those confused. /images/graemlins/confused.gif So you can pattern map the river at Stars, and the turn at Party, but not vice versa, is that it??

[CountDuckula]

[ QUOTE ]
[ QUOTE ]
Read my two posts in this thread, before you're wasting any more time on this holy mission of yours.

[/ QUOTE ]

I know exactly what you're talking about but my calculation is still correct, IMO. You gotta trust me on this. I'm getting around 10% when it should be around 4%. I have taken everything into account. Just wait for more results cuz it's not significant yet. THANK YOU.

[/ QUOTE ]

What you're failing to take into account is the number of people who have pocket pairs and are still present for the river. Each such person adds another 2 outs to the possibility of someone making a set; as was pointed out above, if 2 people have pocket pairs of different ranks the probability that one of them will make a set approximately doubles. And if 3 people have pairs, it approximately triples, compared to the single person holding a pair. It's like the lottery; the chance of any single ticket winning the jackpot may be 50 million to one (depending on the exact game), but if there are 25 million unique tickets out there, the chance that one of them will win the jackpot is about even.

Your 11% figure could well be accounted for by such considerations, even if you do have a statistically significant sample size (how many hands are we talking about, anyway?). He had 2 outs to beat you; you had the other 45 or so -- but would you even be here if you were the one who had rivered the set, or if neither of you had? He got lucky, no doubt. But it happens, even in live games.

-Mike

[AleoMagus]

When you finish this, be sure to calculate what kind of standard deviation you can expect from your 4-5%. After that, calculate a statistical confidence level in your findings.

Just a wild guess would be that after 100,000 hands, you'd still be well within a reasonable expectation with even double the anticipated amount of sets being made on the river.

You might also be interested to know that Kahwanake and other gaming regulators that sign off on all the major sites do the very kinds of calculations that you are attempting now - but with millions of samples. They seem satisfied with the results.

So, if you are trying to prove how unlucky you are, go right ahead. I'll take your word for it. If you are trying to prove that the site is rigged... I'll take kahwanake's (however it is spelled) word that it isn't.

Regards
Brad S

PS - The pattern map was a big joke started by ulysses on this site. If they gave away nobel prizes for humour, he'd get it for this one.