Imbalanced

Discussions about BGA (all languages)
Forum rules
Warning: challenging a moderation in Forum = 10 days ban
More info & details about how to challenge a moderation: viewtopic.php?p=119756
Ceaseless
Posts: 1303
Joined: 12 November 2022, 17:06

Re: Imbalanced

Post by Ceaseless »

Meeplelowda wrote: 05 January 2024, 00:43
Ceaseless wrote: 04 January 2024, 20:03
Cile wrote: 04 January 2024, 16:49The vast majority of tables on BGA are turn based, so you'll be more likely to find an opponent around your skill level.
Wait, seriously? What makes you say that? I'd expect there to have been far more real-time based games on BGA than turn-based. It's a shorter format, easier and more conventional to work with. I don't have anything that would confirm my hypothesis one way or the other though. I could see there being more ongoing turn-based tables since they take forever, so it would just be the ongoing real-time games from that part of the day vs the turn-based ones from not just that day, but any ongoing from that week, month, etc. but I'd be surprised if there were more overall turn-based games in general. Are there stats on this?
Look at this page to see the relative numbers of real-time vs. turn-based games that are currently being played. True, it's not an apple to apples comparison because of course real-time games are shorter, but they'd have to be a lot shorter for their volume to be comparable to the turn-based volume.

Why the difference? Although it's not my personal preference, playing turn-based literally lets you play dozens (or even hundreds) of the same game simultaneously. Even though real-time is "more conventional to work with," I'd argue that turn-based is more convenient to work with. You aren't tied to being in a certain place for an extended period of time (which matters for longer format games). Also, there's no need to arrange for a group of friends to all be available at a certain time (if you don't want to play with random strangers).
Like I noted above, I could see there being more ongoing turn-based games than real-time because you're dealing with a larger slice of time. All of the real-time games you're counting are games from the last few minutes, or maybe an hour if it's a long game. But you're counting turn-based games from say, last week, etc. The real-time games started back then would have all been finished by now. The large group of completed games within that timeframe would likely disproportionately eliminate the real-time games over the turn-based ones. I'd expect it to be rather severe, given the slice of time difference we're talking about (minutes/hour vs days/weeks+). Wonder if there's a way to pull up finished game numbers in a timeframe, or see the rate at which tables tend to successfully start. Also curious to see how it changes across games. (My initial guess would be turn-based seeing more relative success in longer games as well as less popular ones.)

Edits: Have added some things
User avatar
Shivaware
Posts: 322
Joined: 26 November 2012, 17:00

Re: Imbalanced

Post by Shivaware »

Ceaseless wrote: 05 January 2024, 01:36
Meeplelowda wrote: 05 January 2024, 00:43
Ceaseless wrote: 04 January 2024, 20:03

Wait, seriously? What makes you say that? I'd expect there to have been far more real-time based games on BGA than turn-based. It's a shorter format, easier and more conventional to work with. I don't have anything that would confirm my hypothesis one way or the other though. I could see there being more ongoing turn-based tables since they take forever, so it would just be the ongoing real-time games from that part of the day vs the turn-based ones from not just that day, but any ongoing from that week, month, etc. but I'd be surprised if there were more overall turn-based games in general. Are there stats on this?
Look at this page to see the relative numbers of real-time vs. turn-based games that are currently being played. True, it's not an apple to apples comparison because of course real-time games are shorter, but they'd have to be a lot shorter for their volume to be comparable to the turn-based volume.

Why the difference? Although it's not my personal preference, playing turn-based literally lets you play dozens (or even hundreds) of the same game simultaneously. Even though real-time is "more conventional to work with," I'd argue that turn-based is more convenient to work with. You aren't tied to being in a certain place for an extended period of time (which matters for longer format games). Also, there's no need to arrange for a group of friends to all be available at a certain time (if you don't want to play with random strangers).
Like I noted above, I could see there being more ongoing turn-based games than real-time because you're dealing with a larger slice of time. All of the real-time games you're counting are games from the last few minutes, or maybe an hour if it's a long game. But you're counting turn-based games from say, last week, etc. The real-time games started back then would have all been finished by now. The large group of completed games within that timeframe would likely disproportionately eliminate the real-time games over the turn-based ones. I'd expect it to be rather severe, given the slice of time difference we're talking about (minutes/hour vs days/weeks+). Wonder if there's a way to pull up finished game numbers in a timeframe, or see the rate at which tables tend to successfully start.
If you go to the game information section, the number of games played entry is a clickable link that shows recently completed games. Unfortunately, you have to click each one to see what type it is. My limited checking showed there were more real-time games completed. Obviously, this will depend on the time of day. For Through the Ages (New Story), I counted 14 real-time completed games versus 7 turn-based. I predict lighter games will skew even more towards real-time games.
User avatar
Romain672
Posts: 1348
Joined: 05 April 2016, 13:53

Re: Imbalanced

Post by Romain672 »

I checked, here is my results (no difference with random hours it seem):
- 19 real time games
- 2 turn based games
- 11 table leaved before full
- 2 training mode
- 2 'phantom games' (where the number of the table lead nowhere)

So yeah, even if my sample is small, that show a massive difference between real time and turn based.
User avatar
Kayvon
Posts: 1059
Joined: 17 October 2011, 01:39

Re: Imbalanced

Post by Kayvon »

Realtime games are more popular overall. BGA doesn't offer full stats to everyone, but developers can see stats of games they have implemented. Here's a sample of some of the top games I've worked on where I tell you the number of realtime games and the number of "friends only" games, which are not timed and are commonly played as either realtime or turn-based. It's possible "friends only" includes all solo games, in which case those are heavily skewed towards realtime games as well.

Game A: 69% realtime (and an additional 17% in friends only mode)
Game B: 62% realtime (and an additional 11% in friends only mode)
Game C: 55% realtime (and an additional whopping 25% played in friends only mode)

If we remove the friends only mode from the distribution and just sort on realtime vs turn-based, those numbers become 83%, 70%, and 73% realtime tables.

There are far more realtime plays than there are turn-based.
User avatar
ufm
Posts: 2039
Joined: 06 January 2017, 08:38

Re: Imbalanced

Post by ufm »

turtler7 wrote: 04 January 2024, 21:02 The number of turn based games is much much higher than the real time ones for most games.
Completely wrong. As Kayvon said, real-time mode dominates over turn-based mode if the game has a large enough player pool.
The numbers of real-time/turn-based tables in all games can be checked from Audience section of Control panel (available to developers and admins only).
They provide the approximate ratio. Some games (usually trick-taking or hidden role) even have 100 ~ 1000 : 1 of RT : TB ratio.
Stroom
Posts: 428
Joined: 14 July 2016, 19:10

Re: Imbalanced

Post by Stroom »

Meeplelowda wrote: 05 January 2024, 00:43
Ceaseless wrote: 04 January 2024, 20:03
Cile wrote: 04 January 2024, 16:49The vast majority of tables on BGA are turn based, so you'll be more likely to find an opponent around your skill level.
Wait, seriously? What makes you say that? I'd expect there to have been far more real-time based games on BGA than turn-based. It's a shorter format, easier and more conventional to work with. I don't have anything that would confirm my hypothesis one way or the other though. I could see there being more ongoing turn-based tables since they take forever, so it would just be the ongoing real-time games from that part of the day vs the turn-based ones from not just that day, but any ongoing from that week, month, etc. but I'd be surprised if there were more overall turn-based games in general. Are there stats on this?
Look at this page to see the relative numbers of real-time vs. turn-based games that are currently being played. True, it's not an apple to apples comparison because of course real-time games are shorter, but they'd have to be a lot shorter for their volume to be comparable to the turn-based volume.

Why the difference? Although it's not my personal preference, playing turn-based literally lets you play dozens (or even hundreds) of the same game simultaneously. Even though real-time is "more conventional to work with," I'd argue that turn-based is more convenient to work with. You aren't tied to being in a certain place for an extended period of time (which matters for longer format games). Also, there's no need to arrange for a group of friends to all be available at a certain time (if you don't want to play with random strangers).
Let's take Azul.

Currently I see 109 real time games vs 8486 turn based games.

1 real time game takes 12 minutes according to the game page. That's 60/12 = 5 games per hour, 5*24 = 120 games per day. 120 * 109 = 13080 real time games played per day. A very rough average.

Now take 8486 turn based games. How long do those last? I'd imagine that it would take more than 1 day usually. Each player can take quite a long time to respond. In this case, there are definitely more real time games played.

To break even, turn based games would have to last 8486/13080*24 = 15.5 hours on average. I'm sure most turn based games last longer than that.

If there would be a way to find out how much the turn based games usually last, we could compare better.

Ark Nova
245 vs 9711
1 hour = 24 games per day = 5880 real time games per day
Turn based = 9711/5880 = 1.5 days per game or 39.6 hours to break even. I'm sure most ark nova turn based games would take at least a week.
Last edited by Stroom on 05 January 2024, 07:08, edited 1 time in total.
User avatar
ufm
Posts: 2039
Joined: 06 January 2017, 08:38

Re: Imbalanced

Post by ufm »

Stroom wrote: 05 January 2024, 07:01
Correct.

Azul RT-TB in Dec. 2023: 481667 - 34661 (93.3% - 6.7%)
Ark Nova RT-TB in Dec. 2023: 201996 - 62160 (76.5% - 23.5%)

There are even more extreme examples:

Bang! RT-TB in Dec. 2023: 43754 - 466 (98.946% - 1.054%)
French Tarot RT-TB in Dec. 2023: 23703 - 48 (99.8% - 0.2%)

The number of ongoing games cannot represent the number of all tables.
Last edited by ufm on 05 January 2024, 07:15, edited 1 time in total.
Stroom
Posts: 428
Joined: 14 July 2016, 19:10

Re: Imbalanced

Post by Stroom »

ufm wrote: 05 January 2024, 07:06
Stroom wrote: 05 January 2024, 07:01 Let's take Azul.
Correct. Azul RT-TB in Dec. 2023: 481667 - 34661 (93.3% - 6.7%).
The number of ongoing games cannot represent the number of all tables.
I added Ark Nova calculation too. I assume it's the same. How did you find that stat?

There can be quite many "abandoned" games in turn based as well - players taking maybe one turn per day or something. If the game takes on average 20-40 turns, it's going to skew the statistics much more. If it's per player, you could multiply that by 2-3-4 and a game would basically last half a year.
User avatar
ufm
Posts: 2039
Joined: 06 January 2017, 08:38

Re: Imbalanced

Post by ufm »

Stroom wrote: 05 January 2024, 07:13 I added Ark Nova calculation too. I assume it's the same. How did you find that stat?
I mentioned it above:
ufm wrote: 05 January 2024, 05:49 The numbers of real-time/turn-based tables in all games can be checked from Audience section of Control panel (available to developers and admins only).
They provide the approximate ratio. Some games (usually trick-taking or hidden role) even have 100 ~ 1000 : 1 of RT : TB ratio.
Also, added more examples as well.
User avatar
Meeplelowda
Posts: 3833
Joined: 14 March 2020, 10:31

Re: Imbalanced

Post by Meeplelowda »

ufm wrote: 05 January 2024, 07:06
Stroom wrote: 05 January 2024, 07:01
Correct.

Azul RT-TB in Dec. 2023: 481667 - 34661 (93.3% - 6.7%)
Ark Nova RT-TB in Dec. 2023: 201996 - 62160 (76.5% - 23.5%)

There are even more extreme examples:

Bang! RT-TB in Dec. 2023: 43754 - 466 (98.946% - 1.054%)
French Tarot RT-TB in Dec. 2023: 23703 - 48 (99.8% - 0.2%)

The number of ongoing games cannot represent the number of all tables.
That analysis seems reasonable. I stand corrected. Although you'd expect the bias toward RT in a social deduction game like Bang!, that bias is not present in Azul, yet RT still predominates. Ark Nova surprises me a bit. How long does a BGA RT game take vs. over the board?
Play Now lobby 4ever! https://boardgamearena.com/lobby
Post Reply

Return to “Discussions”