Dynasties vs. Parity

OakDragon · 07-10-2006, 04:26 AM

I have found what I think is a bug regarding AI team strategies. Take a look at the league history screen below. Notice how it seems to start out like there are dynasties that the league goes through. One team will win championships for a few years, then another, and so on. After awhile (about 2025 in this screenshot), the championships seem to stop being "streaky"... a new team is champion every year.

I think this change is a bug and I have a hypothesis as to why it happens. I noticed when I started the league that all of the AI teams had random sets of game strategies. Recently, I looked again and noticed the had all been set to mid levels (see Team Strategies screen below). My hypothesis is that when the original AI team managers were fired or retired, the newly generated replacements had everything set to a midline instead of random variations. This led to more similarity in the performance of the teams and thus more random championships.

I suppose it can be argued that the random way is better (and perhaps it should be an option), but to change like this in the middle of a league history as the appearance of being a bug. (I happen to like the dynasty version better myself anyway. Also note that the roster strategies seem to be still random. This may be either because the bug does not apply to those or because the GM either doesn't turn over or does but hasn't yet.)

I'm running this on the Mac version with the 1.0.2 patch.

battists · 07-11-2006, 04:33 PM

Hmm, this is interesting.

I need to dig up some people to do some more legwork on this, but I think we have to come up with something more conclusive if we want Markus' attention on this.

First of all, we're going to need to sim a bunch of 25+ year leagues to see if this pattern bears out, or if it was just your league. I'm not convinced there's anything to the relationship.

For example, I have one simple fictional league that I simmed out for 120+ years, starting in 2006, that I use when looking for long-term or historical reporting issues. Early on there were some dynasties, but it didn't level off.

In the first 25 years of the league:

SF appeared in 3 in a row, and won 2
Fort Worth appeared in 5 in 6 years, winning just once (hello Atlanta!)
Indy won 3 in a row

There were several others that won twice in a row or twice in three years.

But, leaping ahead:

- The Detroit Isotopes (

) won 3 straight from 2033-2035.
- The Fort Worth Hawkeyes appeared in 3 of 4 from 2037-2040, winning 2.
- The Atlanta Daredevils won 3 straight from 2048-2050.
- The Dallas Butterflies were unstoppable in the 2050s, appearing in 8 of 9 championships and winning SEVEN of them.

- The Irving Roadrunners were a force in the 2070s, missing the championship game just twice in the 11 years from 2068-2078, winning 7 of those!

Anyway, that was fun, but my point is, I think a lot more info is needed before I would call the connection relevant. In my year, I saw long spans of both dynasties and parity.

It may be a lot easier to prove simply that new hire managers all have default settings. But, I'm not sure that it ties to dynasties.

Syd Thrift · 07-11-2006, 05:05 PM

How big is your league? Because I may not be seeing dynasties but I am seeing the same 3 teams win in recent years.

battists · 07-11-2006, 05:10 PM

Quote:

Originally Posted by Syd Thrift

How big is your league? Because I may not be seeing dynasties but I am seeing the same 3 teams win in recent years.

16 teams total.

A lot of stuff impacts this as well, not the least of which would be revenue sharing, free agency, and other financial concerns.

Mine was a completely default 16-team fictional league.

OakDragon · 07-11-2006, 06:16 PM

Mine was 16 teams also. It gets weirder, though. I simmed some more years and found that some of the AI's strategies had been randomized again, and some AI's still had everything set to midlevels.

battists · 07-12-2006, 12:49 AM

Quote:

Originally Posted by OakDragon

Mine was 16 teams also. It gets weirder, though. I simmed some more years and found that some of the AI's strategies had been randomized again, and some AI's still had everything set to midlevels.

So their strategies are changing? Kinda cool!

OakDragon · 07-12-2006, 03:16 AM

Yeah, as long as it's meant to be that way. (And as long as I didn't inadvertently do anything to screw things up.) I just thought it was kinda weird (and not good) that they were all set to mid-levels for awhile.

battists · 07-12-2006, 09:54 AM

Oak, at this point, I'm inclined to move this to Closed issues. I'm just not seeing any convincing evidence that a bug exists here.

Izzat OK with you?

Steve

OakDragon · 07-12-2006, 12:52 PM

Quote:

Originally Posted by battists

Oak, at this point, I'm inclined to move this to Closed issues. I'm just not seeing any convincing evidence that a bug exists here.

Izzat OK with you?

Steve

Yeah, that's fine. I'm not sure there is either. For now, it's just weird.

OakDragon · 07-28-2006, 05:33 PM

I found a reason that these issues ought to be looked at again. I noticed that if you go to one of the teams' strategies pages where the strategies are all midlined (12262) and you then go down to the menu at the bottom to ask the manager for team strategies, they are changed. So, in essence, the game is midlining the strategies for "reasons" of its own. IMO, this is not right. The game ought to be getting the strategies from the AI manager. (And this AI manager should come from a pool of personnel in the game who have varied sets of strategies.)

Football23 · 07-30-2006, 06:01 PM

My current MLB league in which I'm in my 12th year, the first year the A's won it, the next year the Blue Jays won it, and my Red Sox have now won 9 straight World Series. (no I'm not cheating, editing, etc.)

Currently it is the 18th of September 2017 and I am 134-15. Most likely the Sox will win 10 in a row this season. I have also noticed that big market teams like the Yankees (using CubbyFan's rosters by the way) fold out early, in the first few years, and never get back to being good which makes it kind of boring when I'm already dominant as the Red Sox plus there isn't even a good Yankee team fielded to spice up the rivalry. (If only that were the case in real life! lol)

Anyways, how is it possible that my team can do this well? Is it just bad AI?

07-10-2006, 04:26 AM	#1
OakDragon Hall Of Famer Join Date: Feb 2006 Location: San Jose, CA USA Posts: 3,494	Dynasties vs. Parity I have found what I think is a bug regarding AI team strategies. Take a look at the league history screen below. Notice how it seems to start out like there are dynasties that the league goes through. One team will win championships for a few years, then another, and so on. After awhile (about 2025 in this screenshot), the championships seem to stop being "streaky"... a new team is champion every year. I think this change is a bug and I have a hypothesis as to why it happens. I noticed when I started the league that all of the AI teams had random sets of game strategies. Recently, I looked again and noticed the had all been set to mid levels (see Team Strategies screen below). My hypothesis is that when the original AI team managers were fired or retired, the newly generated replacements had everything set to a midline instead of random variations. This led to more similarity in the performance of the teams and thus more random championships. I suppose it can be argued that the random way is better (and perhaps it should be an option), but to change like this in the middle of a league history as the appearance of being a bug. (I happen to like the dynasty version better myself anyway. Also note that the roster strategies seem to be still random. This may be either because the bug does not apply to those or because the GM either doesn't turn over or does but hasn't yet.) I'm running this on the Mac version with the 1.0.2 patch. Last edited by OakDragon; 05-21-2007 at 04:13 PM.

07-28-2006, 05:33 PM	#10
OakDragon Hall Of Famer Join Date: Feb 2006 Location: San Jose, CA USA Posts: 3,494	I found a reason that these issues ought to be looked at again. I noticed that if you go to one of the teams' strategies pages where the strategies are all midlined (12262) and you then go down to the menu at the bottom to ask the manager for team strategies, they are changed. So, in essence, the game is midlining the strategies for "reasons" of its own. IMO, this is not right. The game ought to be getting the strategies from the AI manager. (And this AI manager should come from a pool of personnel in the game who have varied sets of strategies.) Last edited by OakDragon; 07-28-2006 at 05:34 PM.

07-11-2006, 04:33 PM	#2
battists Hall Of Famer Join Date: Jul 2004 Posts: 18,506	Hmm, this is interesting. I need to dig up some people to do some more legwork on this, but I think we have to come up with something more conclusive if we want Markus' attention on this. First of all, we're going to need to sim a bunch of 25+ year leagues to see if this pattern bears out, or if it was just your league. I'm not convinced there's anything to the relationship. For example, I have one simple fictional league that I simmed out for 120+ years, starting in 2006, that I use when looking for long-term or historical reporting issues. Early on there were some dynasties, but it didn't level off. In the first 25 years of the league: SF appeared in 3 in a row, and won 2 Fort Worth appeared in 5 in 6 years, winning just once (hello Atlanta!) Indy won 3 in a row There were several others that won twice in a row or twice in three years. But, leaping ahead: - The Detroit Isotopes ( ) won 3 straight from 2033-2035. - The Fort Worth Hawkeyes appeared in 3 of 4 from 2037-2040, winning 2. - The Atlanta Daredevils won 3 straight from 2048-2050. - The Dallas Butterflies were unstoppable in the 2050s, appearing in 8 of 9 championships and winning SEVEN of them. - The Irving Roadrunners were a force in the 2070s, missing the championship game just twice in the 11 years from 2068-2078, winning 7 of those! Anyway, that was fun, but my point is, I think a lot more info is needed before I would call the connection relevant. In my year, I saw long spans of both dynasties and parity. It may be a lot easier to prove simply that new hire managers all have default settings. But, I'm not sure that it ties to dynasties.

07-11-2006, 06:16 PM	#5
OakDragon Hall Of Famer Join Date: Feb 2006 Location: San Jose, CA USA Posts: 3,494	Mine was 16 teams also. It gets weirder, though. I simmed some more years and found that some of the AI's strategies had been randomized again, and some AI's still had everything set to midlevels.

07-12-2006, 03:16 AM	#7
OakDragon Hall Of Famer Join Date: Feb 2006 Location: San Jose, CA USA Posts: 3,494	Yeah, as long as it's meant to be that way. (And as long as I didn't inadvertently do anything to screw things up.) I just thought it was kinda weird (and not good) that they were all set to mid-levels for awhile.

07-12-2006, 09:54 AM	#8
battists Hall Of Famer Join Date: Jul 2004 Posts: 18,506	Oak, at this point, I'm inclined to move this to Closed issues. I'm just not seeing any convincing evidence that a bug exists here. Izzat OK with you? Steve

07-30-2006, 06:01 PM	#11
Football23 Minors (Single A) Join Date: May 2006 Posts: 81	My current MLB league in which I'm in my 12th year, the first year the A's won it, the next year the Blue Jays won it, and my Red Sox have now won 9 straight World Series. (no I'm not cheating, editing, etc.) Currently it is the 18th of September 2017 and I am 134-15. Most likely the Sox will win 10 in a row this season. I have also noticed that big market teams like the Yankees (using CubbyFan's rosters by the way) fold out early, in the first few years, and never get back to being good which makes it kind of boring when I'm already dominant as the Red Sox plus there isn't even a good Yankee team fielded to spice up the rivalry. (If only that were the case in real life! lol) Anyways, how is it possible that my team can do this well? Is it just bad AI?