Game is Unplayable

[azulmofo]

The algorithm is absolutely ridiculous. Game is not playable in current state. I’ve played some terrible backgammon apps but nothing as ridiculously bad as this. The opponent didn’t even know how to play. Had 5 to 6 open at all times. And won. Brutal. And unplayable

[Remkar]

The algorithm is absolutely ridiculous. Game is not playable in current state. I’ve played some terrible backgammon apps but nothing as ridiculously bad as this. The opponent didn’t even know how to play. Had 5 to 6 open at all times. And won. Brutal. And unplayable

Lol.

Your 3 games of backgammon on here definitely make you more of an expert than the many, many thousands of games and numerous statistical analyses that many of us (mathematics, statisticians, scientists, computer programmers, psychologists, expert backgammon players, etc.) have done over the years.

It's one thing to come on here asking questions, but to come on here after 3 games and declare that you've found a problem is pretty disrespectful. And you didn't even take 30 seconds to look at other forum posts where this has been discussed in detail along with supporting tests and probability discussions.

Feel free to delete or edit your post and apologize when time allows.

[euklid314]

The algorithm is absolutely ridiculous. Game is not playable in current state. I’ve played some terrible backgammon apps but nothing as ridiculously bad as this. The opponent didn’t even know how to play. Had 5 to 6 open at all times. And won. Brutal. And unplayable

You are obviously talking about game #439366395.

Yes, your opponent played very unorthodox and surely not the best moves. He forced himself into a backgame from the first move on, by offering you lots of checkers to hit. What you do not seem to know is the fact that this tactic has quite high winning chances. I would not be surprised if you *never* exceeded 70% winning chances - not in a *single* moment of the whole game. Your opponent ended in a very promising endgame with 3 points (1+3+5) made in your homeboard and a good timing when you had to leave shots. I will further analyze, but my feeling is that he had 50% winning chance late in the game, even being 100+ pips behind in the race.

Why was your opponents tactic wrong nevertheless, even if it is often successful?
*) If your opponent loses the backgame he will lose it with gammon most of the time. This argument is not valid on BGA, though, since we play 1-point matches here and there are no gammons.
*) It is very difficult to ensure a good timing in a backgame. I.e., one needs a strong homeboard oneself at exactly the moment when the race-leading player tries to bring his last checker home (and is forced to leave double or triple shots).

Perhaps the opponent of azulmofo likes to play backgames or perhaps he (incorrectly) believes that his strategy is superior. You never know... But letting your opponent hit a lot of checkers and setting up a decent backgame makes an interesting game with very difficult and complex gameplay, for sure!

It is not easy to find the correct moves against such players and I would not be surprised if XG will rate the moves of this game as quite badly played by both players. I will continue with another post as soon as I have analyzed it with XG...

[euklid314]

After my initial thoughts, I have analyzed the game with XG. Many of my assumptions were correct.

*) Both players played suboptimal. XG calls the PR of azulmofo (25,37) "Beginner" and the PR of his opponent (35,80) "Distracted" - which is worse than "Beginner". Note that all backgames are very difficult to play - even top players have bad PR's (Performance Ratings) when playing a back game. There is no shaming involved here - a bad PR is to be expected, only computers play these positions near perfectly.

*) azulmofo was never better than 63% win rate. Most of the time he was even underdog.

I want to demonstrate this with a screenshot that shows a very significant moment in the game.
https://imgur.com/CyqWySD

White (=azulmodos opponent) did roll a 3-1 and he played Bar/22, 6/5 and thus could block azulmodos 3-point, additionally to the 1-point and 5-point which he had secured earlier. A game situation where you have made 2 or more points in your opponents board is called a backgame. The timing is ideal for white since he has a strong board himself (1-,4-,6-points made) and has additional checkers on 9 and 13 such that he does not have to leave his defensive position in the next roll.

White is 53%-favorite to win the position of the screenshot! Even being 97 pips down (brown only needs 87 pips to end the game, white needs 184 pips), white is a favorite! I am quite sure that azulmofo misjudged the situation and thought he would easily win the game from this position - but he was the underdog.

The reason why white is the favorite: Black has lots of forced moves and will almost with guarantee leave shots sooner or later. And then white has not only one number to hit the checker but he will get at least a double shot.

[ArianeKl5]

The algorythm for Backgammon here on BGA seems to be very unfair - I have played many many games and it is ridiculous how many doubles are falling - especially for the other player.

I really like to play backgammon - but here on BGA it is driving me crazy.

Best wishes
Ariane

[Jellby]

it is ridiculous how many doubles are falling - especially for the other player.

... and I'm sure the other player thinks the same

[Serilo]

The algorithm is absolutely ridiculous. Game is not playable in current state. I’ve played some terrible backgammon apps but nothing as ridiculously bad as this. The opponent didn’t even know how to play. Had 5 to 6 open at all times. And won. Brutal. And unplayable

Why was your opponents tactic wrong nevertheless, even if it is often successful?
*) If your opponent loses the backgame he will lose it with gammon most of the time. This argument is not valid on BGA, though, since we play 1-point matches here and there are no gammons.

Well, that is surely the problem with this implementation, then?
If an implementation removes a punishment for a certain playstyle then the implementation isn't faithful to the original(i.e. should not have made it out of alpha, let alone released...)

[Romain672]

Well, that is surely the problem with this implementation, then?
If an implementation removes a punishment for a certain playstyle then the implementation isn't faithful to the original(i.e. should not have made it out of alpha, let alone released...)

This is a problem with bga, not with the implementation.

When you play any board game, being able to say after a game that someone 'won twice' and another player 'loosed twice' in one game is really not commun.
I can think of a single hand of poker where you could argue that you can lose 100 times in one game, but it look pretty different.
I don't have any other game in mind where it could be the case. Maybe the Barbu ( https://en.wikipedia.org/wiki/Barbu_(card_game) ) but again it look different since each big game is separated in four.

But so, since it look like on those game each game would require a lot of attention to change how elo work, bga decided to not do it.

[Eyeleen]

I couldn't agree more. For me it's just the interface. I've played backgammon on a lot of sites and this version is just so small and so slow to move pieces. It's just not pleasant. I would rather play on sites I know aren't particularly 100% fabulous just because the interface is just so much better.

[ufm]

Well, that is surely the problem with this implementation, then?
If an implementation removes a punishment for a certain playstyle then the implementation isn't faithful to the original(i.e. should not have made it out of alpha, let alone released...)

This is a problem with bga, not with the implementation.

In fact, this is a problem of the current implementation as match play is actually the major format of the modern Backgammon.