Teams per fictional division -- 4 vs. 5

tklem321 · 07-07-2018, 11:09 PM

In looking around in these forums I've seen quite a few comments about 5 teams per division being much better than 4 teams per division -- to the point where 4 teams per division is sometimes called an abomination/travesty/etc. However, I never see any reasons given for this. What is the reasoning, if there is any, behind 5 teams being better than 4?

low · 07-07-2018, 11:38 PM

With four teams per division it's more likely for a team with a losing record to win the division.

Personally, I will always prefer having four 4-team divisions per league if I expand, since then I can reward only division winners with playoff births. I'm not a fan of the wild card.

tklem321 · 07-08-2018, 12:04 AM

Thanks for the response. It seems to me that if teams play more games in-division than out, the chances of a team with a losing record winning the division are slim. And, if the teams in that sorry division are that bad, I'm not sure having a 5th team, good or bad, helps a league much.

jimmysthebestcop · 07-08-2018, 01:20 AM

Only league I can think of with 4 team divisions is the NFL and several teams with losing records have won their division and gone to the playoffs since the change happened in recent years.

Think its been around handful of teams since 2010 I know Seahwaks were the 1st team to do it.

But totally personally preference some people its annoying others dont care.

tklem321 · 07-08-2018, 11:31 AM

Thanks. Now I at least know the reasoning. The limited number of games in a season may exacerbate the problem in the NFL. I think I'm going to set up a test league in OOTP and run 20 seasons or so to see what happens.

jimmysthebestcop · 07-09-2018, 03:09 AM

Quote:

Originally Posted by tklem321

Thanks. Now I at least know the reasoning. The limited number of games in a season may exacerbate the problem in the NFL. I think I'm going to set up a test league in OOTP and run 20 seasons or so to see what happens.

Post results so curious as to outcome!

BIG17EASY · 07-09-2018, 01:16 PM

Quote:

Originally Posted by tklem321

Thanks. Now I at least know the reasoning. The limited number of games in a season may exacerbate the problem in the NFL. I think I'm going to set up a test league in OOTP and run 20 seasons or so to see what happens.

I have a league that I started in OOTP15 with the standard MLB setup, then eventually expanded to have four four-team divisions in each league. Only once has a sub-.500 team won a division, and it was in the first year of that set up. I've played about 12-13 seasons since without a recurrence. That's a small sample size, but enough to make me feel comfortable with that alignment.

Westheim · 07-09-2018, 02:36 PM

Meh. I favor 6-team divisions.

tklem321 · 07-09-2018, 09:56 PM

I made 2 test leagues, one with 4 divisions of 5 teams each and one with 4 divisions of 4 teams each. I ran both leagues for 25 years --- 100 division races (4x25).

In the 5-team/division league a division was won by a sub-0.500 team zero times.

In the 4-team/division league a division was won by a sub-0.500 team zero times.

I had wild cards in the 5-team/division league and one sub-0.500 wild card team made the playoffs.

If the goal is to keep sub-0.500 teams out of the playoffs, the number of teams in a division doesn't seem to be matter. What matters is whether there are wild cards and even then it appears to be a rare occurrence.

NoOne · 07-09-2018, 10:21 PM

this is simply about volatility and basics of probability. anything that would reduce volatility should reduce the chances of a crap team winning the division.

Probability:

more teams is likely better... (ceterus paribus, obviously)

let's say it's "10%" chance a team is wretched. 10%^4 vs .10%^5 is a more likely situation. 4 teams is more likely to have a crap team at the top than 5 (.01% vs .001%, if it were truly a 10% chance in this example)

no matter what % you toss in there, this will be true.

Volatility:

# of games vs each other... larger the better. if they do play others, it should be the same 'others' and that should do nothing but help further increase the sample for a more common result.. greater # of games in balance the better. (laymans, not a balanced schedule per se, but that's probably best too, but not needed)

with extreme volatility, odd results are more likely to occur creating greater opportunity for something like a sub-.500 team winning the division.

e.g. think of some team that started off hot then finishes under 500.... bit different but similar in concept. the other 'better' teams have to have terrible luck... fewer games means more likely terrible luck has such a large influence.

Le Grande Orange · 07-10-2018, 12:21 AM

Quote:

Originally Posted by tklem321

I made 2 test leagues, one with 4 divisions of 5 teams each and one with 4 divisions of 4 teams each. I ran both leagues for 25 years --- 100 division races (4x25).

In the 5-team/division league a division was won by a sub-0.500 team zero times.

In the 4-team/division league a division was won by a sub-0.500 team zero times.

The kind of schedule you are using will have an impact. If you are using a balanced schedule, then I would expect the likelihood to go up. If you are using a moderately divisionally-weighted schedule, then probably it'd be less likely.

The main issue with scheduling for a 4x4 arrangement is finding the right balance between division game totals and non-division game totals. There are no good answers to that with 4x4 because of the small number of division opponents and large number of non-division opponents.

Quote:

Originally Posted by Westheim

Meh. I favor 6-team divisions.

Agreed. I think six teams is the perfect number—not too many teams and not too few. It also tends to be easier for scheduling, since with five divisional opponents one gets round numbers, e.g. 70, 80, 90, 100 intradivision games.

dpd376 · 07-11-2018, 01:57 PM

Just as a side note, at the 1994 strike Texas was 10 games or so below .500 and in 1st in mid-August, but they were playing a completely balanced schedule back then.

I would have loved it because I prefered the two division, no wildcard 1969-1993 setup and MLB would've probably gone back to it. But it's personal preference.

Cryomaniac · 07-14-2018, 12:43 PM

Quote:

Originally Posted by Westheim

Meh. I favor 6-team divisions.

A 36 team league is quite an interesting idea.

07-07-2018, 11:09 PM	#1
tklem321 Minors (Triple A) Join Date: Apr 2017 Posts: 246	Teams per fictional division -- 4 vs. 5 In looking around in these forums I've seen quite a few comments about 5 teams per division being much better than 4 teams per division -- to the point where 4 teams per division is sometimes called an abomination/travesty/etc. However, I never see any reasons given for this. What is the reasoning, if there is any, behind 5 teams being better than 4?

07-08-2018, 12:04 AM	#3
tklem321 Minors (Triple A) Join Date: Apr 2017 Posts: 246	Thanks for the response. It seems to me that if teams play more games in-division than out, the chances of a team with a losing record winning the division are slim. And, if the teams in that sorry division are that bad, I'm not sure having a 5th team, good or bad, helps a league much. Last edited by tklem321; 07-08-2018 at 12:05 AM.

07-09-2018, 02:36 PM	#8
Westheim Hall Of Famer Join Date: Apr 2012 Location: Germany Posts: 13,944	Meh. I favor 6-team divisions. __________________ Portland Raccoons, 94 years of excell-.... of baseball: Furballs here! 1983 * 1989 * 1991 * 1992 * 1993 * 1995 * 1996 * 2010 * 2017 * 2018 * 2019 * 2026 * 2028 * 2035 * 2037 * 2044 * 2045 * 2046 * 2047 * 2048 * 2051 * 2054 * 2055 * 2061 1 OSANAI : 2 POWELL : 7 NOMURA \| RAMOS : 8 REECE : 10 BROWN : 15 HALL : 27 FERNANDEZ : 28 CASAS : 31 CARMONA : 32 WEST : 39 TONER : 46 SAITO Resident Mets Cynic - The Mets from 1962 onwards, here.

07-09-2018, 09:56 PM	#9
tklem321 Minors (Triple A) Join Date: Apr 2017 Posts: 246	I made 2 test leagues, one with 4 divisions of 5 teams each and one with 4 divisions of 4 teams each. I ran both leagues for 25 years --- 100 division races (4x25). In the 5-team/division league a division was won by a sub-0.500 team zero times. In the 4-team/division league a division was won by a sub-0.500 team zero times. I had wild cards in the 5-team/division league and one sub-0.500 wild card team made the playoffs. If the goal is to keep sub-0.500 teams out of the playoffs, the number of teams in a division doesn't seem to be matter. What matters is whether there are wild cards and even then it appears to be a rare occurrence. Last edited by tklem321; 07-09-2018 at 09:58 PM.

07-11-2018, 01:57 PM	#12
dpd376 Minors (Double A) Join Date: Dec 2009 Posts: 104	Historic occurance Just as a side note, at the 1994 strike Texas was 10 games or so below .500 and in 1st in mid-August, but they were playing a completely balanced schedule back then. I would have loved it because I prefered the two division, no wildcard 1969-1993 setup and MLB would've probably gone back to it. But it's personal preference. __________________ "I hate astroturf, but I love the game astroturf produces." Bill James

07-07-2018, 11:38 PM	#2
low Hall Of Famer Join Date: May 2010 Location: Wisconsin, USA Posts: 6,207	With four teams per division it's more likely for a team with a losing record to win the division. Personally, I will always prefer having four 4-team divisions per league if I expand, since then I can reward only division winners with playoff births. I'm not a fan of the wild card.

07-08-2018, 01:20 AM	#4
jimmysthebestcop Banned Join Date: Jun 2018 Posts: 1,727 Infractions: 0/2 (5)	Only league I can think of with 4 team divisions is the NFL and several teams with losing records have won their division and gone to the playoffs since the change happened in recent years. Think its been around handful of teams since 2010 I know Seahwaks were the 1st team to do it. But totally personally preference some people its annoying others dont care.

07-08-2018, 11:31 AM	#5
tklem321 Minors (Triple A) Join Date: Apr 2017 Posts: 246	Thanks. Now I at least know the reasoning. The limited number of games in a season may exacerbate the problem in the NFL. I think I'm going to set up a test league in OOTP and run 20 seasons or so to see what happens.

07-09-2018, 10:21 PM	#10
NoOne Banned Join Date: Apr 2015 Posts: 7,273 Infractions: 0/1 (3)	this is simply about volatility and basics of probability. anything that would reduce volatility should reduce the chances of a crap team winning the division. Probability: more teams is likely better... (ceterus paribus, obviously) let's say it's "10%" chance a team is wretched. 10%^4 vs .10%^5 is a more likely situation. 4 teams is more likely to have a crap team at the top than 5 (.01% vs .001%, if it were truly a 10% chance in this example) no matter what % you toss in there, this will be true. Volatility: # of games vs each other... larger the better. if they do play others, it should be the same 'others' and that should do nothing but help further increase the sample for a more common result.. greater # of games in balance the better. (laymans, not a balanced schedule per se, but that's probably best too, but not needed) with extreme volatility, odd results are more likely to occur creating greater opportunity for something like a sub-.500 team winning the division. e.g. think of some team that started off hot then finishes under 500.... bit different but similar in concept. the other 'better' teams have to have terrible luck... fewer games means more likely terrible luck has such a large influence.