dice not right

Post Reply
User avatar
oh oh oh santa
Posts: 6
Joined: Mon Dec 06, 2010 2:34 pm

dice not right

Post by oh oh oh santa » Wed Dec 08, 2010 11:26 am

If the dice are ok in backgammon, then why do I get 6 doubles in a row? This happens more then it should in a game, Also when u throw a 6-3 the other guy gets a three on one of his dice almost all the time. Something is wrong with the dice , this would not happen in a normal game, it would really be rare for that many dbles in any game. I have played games where the other person threw mostly doubles to win the game. OF course they say there is nothing wrong with this picture.

User avatar
Sailing_Away
Guide
Posts: 878
Joined: Fri Feb 08, 2008 10:44 am
Location: Perth, ON

Re: dice not right

Post by Sailing_Away » Wed Dec 08, 2010 11:57 am

The perception that there must be something wrong just because the improbable happens is common on sites like this, but such an observation does not serve as proof that the randomization of dice or cards is at fault.

The odds of rolling a double per turn is 1 in 6. Rolling doubles six times in a row is 1 in 46,656. While those seem like wild odds, they really aren't. They are about the same as driving a hole in one in golf (1 in 45,000). People win lotteries where odds exceed 2,000,000 to 1.

In Roulette, the same colour can come up spin after spin after spin..... 16 times in a row would have the odds of 1 in 65,536. Would you believe people report as many as 26 times in a row the same colour has hit? The odds: 1 in 67,108,864.

When dealing with statistics and probability, you cannot ever discount something that happened as being legitimate because it was improbable. Anything that is possible can happen, and six consecutive doubles in backgammon is not hard to believe.
-Brian

"Come Sail Away with me...."

Scatoose
Posts: 5
Joined: Tue Jan 11, 2011 4:55 pm

Re: dice not right

Post by Scatoose » Tue Jan 11, 2011 5:41 pm

Santa's misgivings about Hardwood's random dice generator are for the wrong reasons.
Additionally, "Sailing_Away's" response is typical for servers with cooked generators, not to mention his faulty
explantion of the laws of odds & probabilities.

It's not doubles that should be questioned. It's the uncommon occurance of high odds rolls, not dbls.

The random sample I've used on Hardwood's backgammon server is 50 games. That's approx. 1000 rolls on each side.
Example: if you need a roll of 6-3 to enter a 5/6 closed board and escape to hit an open blot, your odds are 18-1; 6-3, 3-6.
The odds are the 2nd highest possible at a whopping 5.56 in 100 of success. Odds that $$$ players sell their children to attain.
My stats show that in 250 + similar incidents these rolls occur 37.5% of the time. That is 37 times in 100. No way man.

Your explanaton and reasoning is weak. Extremely. The die occurances not only contradict logical odds measurements,
but frequently happen at perfect times during end games, turning the tide of the game in a flash. This happens sure. But I intend to disprove this
flimsy excuse and publish data over 500 and 1000 games.

I've played backgammon for $$$ since high school, 1966. Rolls like these on live boards would prompt homicidal responses.
Don't believe me. Try playing for a couple weeks at your local pub. Then publish your findings. If $$$ are required to play then you have serious
problems if you cook the generator without saying it. Pogo does. F.I.B.S. does also. They maintain that it's required to maintain overall parity
in the server, or else you'd have two distinct cliques (classes) of players: Experts and Novices. Eventually the Novices would bail. Most wouldn't hang
to get beaten to death time after time. If you doubt me, then you're free to look up Yaz at Duffy's in Malden, Mass. Anytime. Bring $100-$500 and I'd be
happy to show you what happens on a live board. Just don't bring rent or baby milk $$$.

I don't know what you call this Hardwood game? But it ain't Backgammon.

Scatoose

Ps. and don't try the certified by experts song either. that doesn't fly on a server that is cooked to begin with.

User avatar
Sailing_Away
Guide
Posts: 878
Joined: Fri Feb 08, 2008 10:44 am
Location: Perth, ON

Re: dice not right

Post by Sailing_Away » Thu Jan 13, 2011 9:36 am

Please do gather data.... lots and lots of data. Streaks and allegoric cases tend to defy logic and odds, but as the number of cases runs out towards infinity, these all average out and the actual results draw closer in line with the mathematical probabilities.

I don't know what reason SCE would have for "cooking" their random generators for dice rolls or card deals. Your tone seems to allege deceitful intent, and I fail to see what motivation they would have for it. A "cooked" algorithm for skewing results would not lead to games that are more fun and would not work towards retaining their players. And if they skewed results so the games were always close, everyone's rating would be huddled right around 1500. It seems apparent that the luck is fairly evenly spread, and it's the skill level that determines over the long haul who the better players are.

I await your results once you have compared actual percentages versus mathematical probabilities with the very large sample size close to infinity.
-Brian

"Come Sail Away with me...."

Scatoose
Posts: 5
Joined: Tue Jan 11, 2011 4:55 pm

Re: dice not right

Post by Scatoose » Sat Jan 22, 2011 3:43 am

How much data would it take? Focus on reality not my tone. Ask anyone who's played 100 + games on the SCE bacgammon server. I've been playing backgammon for $$$ on live boards since 1967. It doesn't take a statistical professor with mounds of data to see the difference with online boards. In fact most online servers are the same.
And some such as Pogo, F.I.B.S., Partouche, Game Colony, et. al, admit they cook their generators.

("I don't know what reason SCE would have for "cooking" their random generators for dice rolls or card deals"?) How about money... it costs to play on SCE. With truely random boards you'd be swimming in cash. It's easy to conceal (after certification). The above servers say a truely random generator would drive away the weaker players, reducing revenues. So don't preach about how I'm whinning about outcomes. Your "tone" seems to imply sour grapes. How many 18-1 and 9-1 miracles do I need to see to convince me what I smell is a rat, not impartiality. Forget doubles. Doubles are a 6-1 proposition. It's the constant in-direct hits and escapes that make it unbelievable.

Sneak into the BG rooms yourself. Play 50 games and tell me if the perfectly timed die rolls affecting outcomes don't dazzle you. Then you bring your cash up to Malden,
MA. once a week and play real live boards and see how it pales. I see scores of excellent players in the Hardwood rooms. Odds players for sure. At the end of games I always say that the games don't seem real and the lion's share agree with me.

Until "you" personally can actually go into the rooms yourself and play 100 games, then come to me and say there's nothing wrong with the randomizer, you'll never know
for sure. Despite the data. You have my handle "Scatoose" and email: Scatoose@gmail.com. Come and play me anytime. My record with this handle is 55-38. That's a
1.45:1 ratio. ???? I'd be ashamed to show my face in town losing 38 games out of 93!!

This server is well designed and laid out. Easy to use and reliable. It's a shame to let it be corrupted with such obvious outcomes. Backgammon is immensely popular.
Just go and visit Play65, BGroom or Gammon Village. They have hundreds of players online constantly. Experts and novices worldwide. Their setups aren't
nearly as well concieved as SCE. Hardwood BG never has more than 30-35 players logged in? Come on man...

User avatar
Sailing_Away
Guide
Posts: 878
Joined: Fri Feb 08, 2008 10:44 am
Location: Perth, ON

Re: dice not right

Post by Sailing_Away » Sat Jan 22, 2011 8:04 am

To recap, what you are saying is not that the randomizer itself is invalid leading to unbelievable streaks, but rather that after the random dice roll or instead of it, SCE has added logic that says if a player needs a certain roll to get off the bar or to win, ignore the random roll and give them the one needed at least enough of the time to exceed what one would expect. Correct me if I am wrong in re-wording your complaint.

I won't discount your sense of "it doesn't feel right" if you are an experienced player. I would probably sense the same thing on a blackjack site if the correct play led to the wrong outcome consistently. It would take lots of playing to get to that point, as it's all probability which doesn't guarantee anything as a certain outcome, but after a while you get the sense that the dealer is a "mechanic". At least at a table you know the dealer changed, and if that dealer follows you to whatever table you go play at, it's time to leave the casino.

This is what you are saying is the point you are at with Hardwood Backgammon, which I have to respect your perspective.
-Brian

"Come Sail Away with me...."

Scatoose
Posts: 5
Joined: Tue Jan 11, 2011 4:55 pm

Re: dice not right

Post by Scatoose » Sat Jan 22, 2011 1:33 pm

So if you agree with my contention that I know what I'm seeing, then my point is taken. An experienced player profiles everything. Including die rolls on live boards as well. I've seen boned dice used on live boards, sure. But when I see 18-1 shots (5.56 chances in 100) happen at least 30% of the time regularly, it just doesn't make sense.
How long do you think Casinos would stay in business if players could bet on any hardway bet and win 30% numbers. No way. The laws of probability are written in stone.
Hot streaks or not. Streaks end. Not at Hardwood. BG players know the real allure of the game; it's managing stiff rolls. Not mechanically moving stones with hot rolls. It's the juice, the action that hooks a gambler in any game.

SCE surely didn't write the code for these modules. Someone did. If the programmer's product chased the majority of players (novice casual players), then SCE surely wouldn find another client package. Otherwise, you'd have two distinct cliques of players: excellent and weak. The weak will go somewhere they can win.

The majority of Hardwood players are below average at best, but still have 50-50 win rates or close to it. Weak players rely on miracles. Not expertise. Losing to good
players is expected. But when a very weak can beat me in a 5 or 9 pt match is really pushing it.

If you want to play, let me know.

User avatar
Jonas
Illuminated One
Illuminated One
Posts: 5035
Joined: Mon Nov 25, 2002 3:46 pm
Contact:

Re: dice not right

Post by Jonas » Sat Jan 22, 2011 5:02 pm

The AI and the game state are different code that the dice roll generator.

There is no logic to influence the dice to be anything other than random. It would probably behoove us to not make it random so folks would be happier, but being purist, and thing things should be what they are, we've been fairly resistant to the idea of "smoothing" out the rolls.

Just to put the rumor to bed (yeah right), there isn't any sort of concept where a player AI or otherwise is getting a diceroll to suit the game state. In fact the roll generator doesn't have a clue who will end up getting the dice.

IF we do anything, it would probably be to modify rolls to match the odds within a smaller set of rolls; we know they look fine in large numbers but of course naturally you get more noise in small numbers of rolls. Its gonna take a lot of work to make that solid and check for bias so we haven't gone down that path yet. The way it would likely work though is something to the effect of looking at a roll, and looking at the last 8 rolls back or something, and look at the current roll and try to make a determination of the chances of getting sequence we're beyond some certain threshold, and roll again (randomly of course). But currently we don't do anything like this, and if we did it would probably be a rule option.

Scatoose
Posts: 5
Joined: Tue Jan 11, 2011 4:55 pm

Re: dice not right

Post by Scatoose » Sun Jan 23, 2011 10:11 am

LOL, What?

I've heard some great razzmatazz and gobblygook in my time but that takes the cake. Whatever you're trying to say, just say it for Christ's sake.

The plain simple facts are that over time, short and long, long odds rolls occur far too often. Timely escapes and bailouts occur at least 30% of the time
and I wouldn't be surprised to see other date increasing that. You need to talk to your players, high and low ratings. If you decide our concerns are real
then fix it. Skip the jive talk and face it Jonas, it's not real.... It's not backgammon

Scatoose
Posts: 5
Joined: Tue Jan 11, 2011 4:55 pm

Re: dice not right

Post by Scatoose » Mon Jan 24, 2011 7:07 pm

Jonas, as a developer of Hardwood Backgammon or the client SCE, this should be published.

Back in 1994-95, F.I.B.S. the (First Internet Backgammon Server) developed the first online competitive backgammon games.
It presented online competition for head to head BG matches. However, a hidden motive existed.

The server was designed and the software coded by a group of programmers, mathematicians, psychologists and world BG experts. The purpose wasn't entertainment,
but a study of human emotional response. The developers wanted to observe human responses in "non-risk games" when the players were given cooked die rolls based
on the current state of each game.

The BG experts consulting on it's developement knew that experienced players would spot contrived outcomes in a short period of time. Since F.I.B.S. was the only
online BG server available, players would return to play despite the obvious bias. But, that wasn't the real focus of this server. The psychologists wanted to study
the emotional reactions of people when bias in their games were seen. Devious to say the least. So goes the fate of a lab guinea pig.

Their findings astounded them. Players would send email and forum posts to the Mods excoriating them for obviously cooking the die rolls. Human responses ranged
from total rage to unbridled arrogance. These findings were published. Years ago I had the link and could have sent it to you; a paper on the entire study. But, I lost
it long ago. If you were to contact the administrator there, he may send it to you.

Now of course you may already know of this study. Therefore, this post is moot and you may have me disappeared. LOL. So you will explain it away and stick to your
insistence on the absolute integrity of your generator. How could you admit it? Everyone would surely believe this site is just another Pogo or Game Colony, admittedly
cooking their generators to boost excitement and membership. But, hardly real Backgammon.

If you are not aware of this study and will admit that the generator solves the winning rolls for a lower odds player or against a player with a large rating or long
winning streak. I believe the latter. I will believe you don't know it's hidden agenda and won't acuse you without proof.

Fact is, the international experts who swarm a site like Play65 won't come near a site like Hardwood BG. Check me out. Your cards rooms are stuffed with players.
BG isn't. Play65, Gammonvillage, Partouche, BGroom are bulging with players; $$$ players. They're convinced those servers use truely random generators.

My suggestion: ditch this program and install a reliable model. Why not. Sitting through games where weak players make repeated critical mistakes and win nearly
half the time just isn't connected to reality. I surmise you won't, due to cost. Designing a new Beta BG site with visually appealing features isn't the answer. Hardwood
is fine the way it is now. In fact I think it's pretty awesome.

Conclusion: Don't claim over and over again that the die rolls are completely un-biased and random and he who complains suffers the typical "sour grapes" syndrome.
It was done with infant technology 15 years ago. If you tell us that it's cooked and live with it, ok... we can... or not. But telling us that it's fair when it's not is bogus.

I invite you to play me anytime and we can discuss what is shown. Don't worry, Hardwood is a free site. You won't need to pay me..... lol

User avatar
Jonas
Illuminated One
Illuminated One
Posts: 5035
Joined: Mon Nov 25, 2002 3:46 pm
Contact:

Re: dice not right

Post by Jonas » Wed Jan 26, 2011 11:32 am

Scatoose wrote:LOL, What?

I've heard some great razzmatazz and gobblygook in my time but that takes the cake. Whatever you're trying to say, just say it for Christ's sake.

The plain simple facts are that over time, short and long, long odds rolls occur far too often. Timely escapes and bailouts occur at least 30% of the time
and I wouldn't be surprised to see other date increasing that. You need to talk to your players, high and low ratings. If you decide our concerns are real
then fix it. Skip the jive talk and face it Jonas, it's not real.... It's not backgammon

haha, Well when I tell the forum its not real, that doesn't seem to help either, hence all the explanation.

We've tested it and there is no significant deviation from the expected statical rolls. We do absolutely no tampering or rigging either. So here we go, the answer is......

We roll the dice, you get what you get which is unbiased. Period.


Hows that for brevity Scatoose?

Might not be the answer you want, but it is what it is.


The only thing left isn't "sour grapes" but a perceptual issue where our minds don't pay attention to the mundane. But it tries to make stuff out of nothing, animals in the clouds for example. In this case, you notice rolls that are interesting, ignoring when things we're boring. You're likely taking the interesting stuff, and trying to create a reason for it. (biased roll to manipulate the game in your case).

Its just not so. I don't know how else to tell you.

Post Reply