After over 500 games I am done with the BGA version

[SherlockC]

I love Catan, before playing on BGA I had played over 1000 games on Colonist, but had got fed up with both bots and some abusive players. I was really happy with the BGA version, and improved to reach the top 100 last October. I continued to rise up the ranks and was delighted to reach number 1 in late November, with an ELO of 478 at one point.
My decline since then has been dramatic, and I don't believe that this is down (entirely) to poor play. I have got more and more frustrated by both crazy dice rolls and continually being third or fourth pick, and begun to question the fairness of this version. Other threads have suggested that the lowest ranked players have an advantage - with their numbers being rolled more frequently than probability suggests. I can accept that the random number generator is broken, but would hate to discover that it is programmed to be biased.
Whatever the reason, I cannot play on BGA any more. I have tried going back to Colonist, but the same problems still persist.
I would love to know what other experienced players think about this.

[Jellby]

Of course, when you raised to #1, randomness had nothing to do with it...

[thoun]

As a game dev, we ask BGA a random number from 1 to 6 when we need a dice roll. There is no player/elo/ranking/date of birth/... parameter to generate a random number. Every conspiracy theory about this is based on human bias towards randomness.

[SherlockC]

Of course, when you raised to #1, randomness had nothing to do with it...

I thought not at the time, but now I am doubting it!

[Stroom]

I looked through some of your games. I'd say that your playstyle itself is not that good.

You often place the settlements in a way that doubles down on the same number (also... why is the random map generator placing the same numbers next to each other? Not using the spiral method for creating games?). This means that your games are going to be more random. If your numbers don't come up as much, the other players will do better. But if they do come up more... you might win but you might as well more often end up in a situation where you have a lot of resources but nothing to do with them.

And the settlements you place... You often place the road towards a location that is almost definitely going to be blocked.

[SherlockC]

I looked through some of your games. I'd say that your playstyle itself is not that good.

You often place the settlements in a way that doubles down on the same number (also... why is the random map generator placing the same numbers next to each other? Not using the spiral method for creating games?). This means that your games are going to be more random. If your numbers don't come up as much, the other players will do better. But if they do come up more... you might win but you might as well more often end up in a situation where you have a lot of resources but nothing to do with them.

And the settlements you place... You often place the road towards a location that is almost definitely going to be blocked.

I certainly don't try to double up numbers, but do not shy away from this if choice of the second settlement is obvious. I accept that in recent games, as my frustration has grown, my settlement choices have been questionable, but again have felt that options were limited, and that taking a chance might give me a hope of the win.

... and, if my playstyle is not that good, doesn't it suggest some weakness in the BGA version if I could get to be the top ranked player!

[Stroom]

Catan itself is quite a random game of itself. You do need some skill but nearly everyone can get high enough that the luck of the dice will start to influence who actually wins in the end. I'd say it's about 30-40% luck based game. No need to take elo seriously and just play for fun.

[DreamerPKU]

I have to agree with SherlockC, and recently I often joked in the chat that the dice must dislike me.

Here is my recent experience with CANTA on BGA. It took 13 games for my ELO to drop from 251 to 103. Frustratingly enough, I was always in the last two players (which explains the ELO decline). Though I confess that there might be something wrong with my tactics, I still think the statistics from several of these games are dubious and partly illustrate why there was a slim chance for me to win.

Game #1 (3p): https://boardgamearena.com/table?table=330276576

In this game, only 5% of rolls were 6 (3 times in total), and 23% were 8 (14 times in total). What's worse, I could hardly afford to buy development cards throughout the game due to either lack of wool (the wool tile numbered 6 produced only 3 times!) or being robbed hard. I was robbed of 10 resources while the other two were robbed of 6 and 1, respectively. I tried my best to get to 7 pts eventually but I still ranked last.

Game #2 (3p): https://boardgamearena.com/table?table=333448373

In this game, I relied on 3 and 10 for lumber. From my past experience, I expected there would be an acceptable (not necessarily sufficient) supply of wood. However, 3 never appeared and 10 only did it twice! Other players relied on wood tiles numbered 4 or 5, which just met their personal needs, so they were unwilling to trade wood. Besides, I bought 4 cards, 2 of which, to my disappointment, were VPs.

There are some other games where the top player prevails and virtually no one can stop she/he from winning (thanks to the dice, robber, and cards again). Such games typically end with the first player getting 10 while others got 4 or 5 at most. Examples include https://boardgamearena.com/table?table=336563449 and https://boardgamearena.com/table?table=336674013 (I got 7 in this game solely because of the largest army special card and a VP card).

Note that all the 4 games aforementioned belong to the 13 most recent games I played on BGA. Though sometimes a player may inadvertently (or deliberately) help another to win, luck can still overwhelms strategic planning.

To conclude, I think the dice on BGA are either biased or compromised. I did well in probability and stochastic process classes, so I know what the LLN has to say. A game typically has about 60 rolls, which usually suffices for LLN to engage for two dice with an entropy of 3.2744.

[Sparon]

continually being third or fourth pick

That is my favorite position to play from. When you are 1st or 2nd, you get an excellent 1st settlement, but your 2nd will be picking from the scraps. In 3rd or 4th position, you can formulate your entire early game strategy pretty effectively with little risk of being blocked off.

[Romain672]

I planned to answer you with stats,but the claims look too chaostic and not enough precise.
The thing is, you compare many differents things in all your games, if you add the probability to have too many/too few of any number between 2 or 12, you add probability for a specific player to have those rolls, add the probability to have X of that number in a row, probability for you to always start 1st/2nd/3rd, having the number you played on turn 1 rolled too many/too few... You end with a pretty high probability.

There was lots of claims in the forum, and it's rare to get anything weird. On the contrary, you find people:
- claiming in each poker website that the game is rigged (while people earn their life with those, so would be ready to pay people to do stats and attack the website)
- claiming in league of legends that the winning queue/looser queue exist
- in among us, that X is too much impostor in a row

There was studies showing that the human is bad about perceiving randomness. So all that seem to lead that bga's randomness is fine.