2012 League Discussion

[Zhukov44]

@DarthMaximus:

IMO, it is very hard to get to 9, 10 wins, even if you try and cherry pick opponents.  It is also hard to not play people.  You still need to post in the looking for a game thread, and anyone can respond.  Based on the standings I just posted, I find it hard to believe that none of the top 6 have played each other.  I of course know for a fact that many of the top 6 have played each other.  And don’t forget there is still the playoff.  The bottom line is you are going to have to play and beat “good” players at some point.

Fair enough, and I agree simplicity is a virtue.

But note that DD and Gamer and I all avoided each other this year.  It wasn’t intentional, but more likely unconscious.  We all hate losing and we all tend to stall when we’re down and we’re all conscious of how playing weaker opponents would advance us in the league.  Had we all played each other, our W/L records might look more like Bold’s record.  Not to say I would have ducked them had they challenged me, but I didn’t challenge them either.  I suspect both of them would agree there’s something wrong with this pattern.  Granting extra points for strength of schedule might help motivate us to challenge each other.  And maybe Gamer or another math expert could help do the calculations for you and JWW.

That said I personally would not grant points for games played alone (I’d prefer the different divisions for that), but do think pure strength of schedule should factor in.

I could see a scenerio where we go back to Major and Minor Leagues.  As Zhuk mentioned, we’d do something like 8-12 for the minor, and 12+ for major.

Good ideas.  I’d bump the minimum to at least 14-15 for the 2nd category.  And maybe a special achievement award for the player with the best record over 20 games.

A deadline for games might be considered, but there will be no judgmet system.  Niether JWW or I want to try and figure out who is winning some 15 rd game.

I’d only advocate judgments for games where it’s obvious who is winning.  Ordinarily this can be done by analyzing the TUV numbers, unless there are special circumstances like a capitol is about to fall or something.  In general if one side has a 25-30% overall tuv advantage then the game is basically over, barring special circumstances where a reasonable case could be made that the other side can come back.  Anyway the reason clear rules and adjudications are needed is to prevent stalling a game one is losing to make the play-off at the expense of someone else.  If the rule is there, most likely very few (if any) games would actually require adjudication.

[gamerman01]

@Zhukov44:

But note that DD and Gamer and I all avoided each other this year.  It wasn’t intentional, but more likely unconscious.  We all hate losing and we all tend to stall when we’re down and we’re all conscious of how playing weaker opponents would advance us in the league.  Had we all played each other, our W/L records might look more like Bold’s record.  Not to say I would have ducked them had they challenged me, but I didn’t challenge them either.  I suspect both of them would agree there’s something wrong with this pattern.  Granting extra points for strength of schedule might help motivate us to challenge each other.  And maybe Gamer or another math expert could help do the calculations for you and JWW.

That said I personally would not grant points for games played alone (I’d prefer the different divisions for that), but do think pure strength of schedule should factor in.

I agree with all of this.

Extra points for strength of schedule would actually not give me incentive to choose different opponents, but that’s OK.  A strength of schedule calc would still help give a more accurate indication of a player’s skills than straight W/L record where everyone’s picking their own opponents and can refuse games.  I should play with some SOS methods on our current standings and see how it changes the order.  I’ll let you know what I find, and what my method is.  Just for fun.

I’m fine without participating in any post-season playoffs, although when I qualify, I’m game to play them for fun.  I enjoy the league play because it’s a good way to get a choice of regular players to challenge, and a good way to get into good, fun A&A games.

[DarthMaximus]

But it can already be a disadvantage to not play potential playoff rivals.  Had you played Dutch or Gamer and beaten them then they would have another loss and you in turn could have those wins and would probably be in the playoffs.  The same could be said for if Gamer or Dutch didn’t play possible playoff rivals.  By not playing potential playoff players (either consciously or unconsciously) you are leaving your playoff hopes in the hands of other players.

Its neither here nor there I guess, cause you can’t play everyone, but it is a huge two game swing if you play and beat a potential playoff rival.  Not only do you get the win, but they get a loss.  IMO, its worth the risk of say a “tougher” game.

–--------------------

Side note.  I did a quick tally using the old Revised scoring system. (based off of Gamers spreadsheet).

Players ranked from Gamer to CTS.  A win over Gamer is currently worth 32 pts a win over CTS is worth 1 pt.  Other notables:

Gamer - 32
LL - 31
Tyz - 30
Zuck - 28
Dutch - 28
JWW - 27
RD - 26
Bold - 25
OBG - 24
etc…
A44 - 3
Ranger - 1
CTS - 1

You also lose 1 pt for every loss.  “Final” score:

Bold 246 - 11 = 235 pts
Gamer 192 - 1 = 191
OBG 183 - 9 = 174
Dutch 154- 2 = 152
LL 149 - 1 = 148
JWW 137 - 3 = 134
Zuck 114 - 2 = 112
RD 104 - 3 = 101

I don’t want to make losses too punitive but even if we go to minus 2 for loss its
Bold = 224
Gamer = 190
OBG = 165
Dutch = 150
LL = 147
JWW = 131
Zuch = 110
RD = 98

No change.  I wouldn’t drop below 3pts per loss.

Now if we did major and minor League that would take Bold and OBG into their own playoff since they have the most games.

But that still leaves a list of:
Gamer - 190
Dutch - 150
LL - 147
JWW - 131
Zuck - 110
RD - 98

Now it took me about an hour to scribble this stuff down (since I’m at work) and I only did the top 8 including Bold and OBG.

Now maybe JWW is gonna get “screwed” out of the playoffs cause Zuck has a better record (at the momemt) but to complete the tally for the remaining members of the league still takes time and I’d estimate another hour or two.  And that was with Gamer doing the work with his spreadsheet and already listing who beat who.

Its hard to justify that time commitment when straight up win % basically handed us the same results in a fraction of the time.

I actually didn’t mind doing the work this morning, b/c it bears out the same results we got with Revised in '07 and '08.
I’m finding a hard time believing that you are going to get drastically different results with an alternate scoring system compared with straight up win %.
I’m still listening to alt scoring methods, but once you hit 8, 10 games played win % seems pretty accurate.  Again for the time commitment we are willing to make.

[DarthMaximus]

One further breakdown would be to make it pts per game:

Gamer - 190/13 = 14.6
Dutch - 150/10 = 15
LL - 147/8 = 18.375
JWW - 131/12 = 10.9
Zuck - 110/10 = 11
RD - 98/9 = 10.9

That would give us:

LL
Dutch
Gamer
Zuck
JWW
RD

Basically the same top 4.  Of course LL needs to get to 10 games.

But pts per game played might be the most accurate.  
(Bold = 8.6, OBG = 8.7)

Edit:

The issue with this method is it becomes risky to play more and more games, and we want to ecourage people to play as many games as possible, not to get to the minumum with a good pt per game ratio and then sit out.

[gamerman01]

Good stuff, Darth.  Yes, your method of points per win and only subtracting one or two per loss rewarded people too much who played a ton of games (Bold, OBG).

Good points about win/loss and who plays who.
I have 20 minutes right now - let me do a quick strength of schedule calc. here……

[gamerman01]

I divided the standings into 3rds by win %.  Top 3rd going through JWW (75%), 2nd 3rd through Journalist (40%).

I counted 3 points for a win against top tier, and -1 for a loss against top tier.
I counted 2 points for a win against middle tier, and -2 for a loss against middle tier.
I counted 1 point for a win against bottom tier, and -3 for a loss to bottom tier.

The results are interesting.

Gamer          17
Boldfresh      16
DutchmanD    11
Lucky            10
Tyzoq              8
JWW              8
Zhukov            7
Dragon            5
SouL                3
OBG                2
Billy                1
Anchovy          1
SAS                -1
Akreider          -1
Journalist          -2

This system’s weakness is that it does not yet account for players who only played 1 or 2 games in the league.

But it is very interesting.  Boldfresh likes to play a lot of games against high quality opponents, to improve his game.  7 of his 11 losses are to top tier players.  This is reflected in my strength of schedule standings.  I would argue that this simple calculation (done in less than 20 minutes) does show you who your top 4 league players of the year were.

Myself, Boldfresh, Dutchman and Lucky.  So a derivation of this system is what I’m recommending for next year.  I would love to hear any arguments that these 4 are not the top 4 players from this year.  :-)

[gamerman01]

btw I’m struggling with Darth and Lucky, but if I were to lose to both of them in 2011, it would change my score to 13.  But under the current system it would drop me to 80% winning percentage and I may not even make the cut for top 4.

The main flaw of the current percentage of win system is that it severely penalizes players like Bold, and rewards players like me and Zhuk.  Bold’s win percentage belies his experience and skills, mainly because he plays on average better players than I do.

[gamerman01]

Made a counting mistake.
Bold would have 8 points, not 16.  So there would be a 3-way tie for 4th: Bold, Tyzoq, and JWW.

[gamerman01]

Note that Bold and Blood and Guts look similar by record and win percentage, but Bold has played significantly stronger opposition than OBG.  The SOS calculation reflects this quite well (for example).  Bold would have 8 points, OBG only 2.  Look at their results on my spreadsheet.  Also, note that in head to head, Bold won 2 out of 3 over OBG.

[DarthMaximus]

Interesting stuff Gamer.

I hadn’t thought about breaking the point totals into tiers.  I’ll have to take a look at it.

If we went with something like this, maybe a simple break of the tiers like this:
70%+
30-70%
Under 30 percent.

Or just making it a 4 tier system with breaks at every 25%.
75-100
50-74
25-49
0-24

The one thing about any type of pt system is you don’t get an official tally until the year is over sice players can go up and down in “ranking”.

In season updates would probably be just listed by win %.

I’ll have to think about this a little more.

[gamerman01]

@DarthMaximus:

The one thing about any type of pt system is you don’t get an official tally until the year is over sice players can go up and down in “ranking”.

Right.  But it takes care of itself in the end.

In season updates would probably be just listed by win %.

Yep

I’ll have to think about this a little more.

Cool

[gamerman01]

@Gamerman01:

Gamer           17
DutchmanD    11
Lucky             10
Boldfresh        8
Tyzoq              8
JWW               8
Zhukov            7
Dragon            3
SouL                3
Billy                3
OBG                 2
Anchovy           1
SAS                 -1
Akreider           -1
Journalist          -2

Fixed Bold and updated for RD vs. Billy result

[DarthMaximus]

Here’s what a 4 tier looks like.

75-100: 4 pts win, 1 pt loss
50-74:  3-2
25-49:  2-3
0-24:    1-4

Gamer: 21
Bold: 18
LL : 17
DD: 15
JWW: 14
Zuck: 10
RD: 9
OBG: 6

I think I did the math right and I didn’t include the RD-Billy game.  Kinda interesting to see.

BTW, Gamer in your spreadsheet, does OBG only have 8 losses.  I may have missed something but it looked like the losses were: LL, DD, RD, Soul (2), Bold (2), Billy.

[666]

A few potential concerns I have with the most recently suggested scoring system.

1. The newly proposed weighted system suggests utilizing the current winning percentage system as a means to assign point values to wins and losses.

I would argue, that if the current percentage based scoring system is deemed inequitable, then, any future system, using the percentage system as its foundation, would be similarly faulted.

An example of this can be realized by reviewing the rankings change of BoldFresh in this proposed system. BF’s winning percentage record regulates a specific value associated with beating him to “X”, and everyone that defeated him get’s a multiple of “X” to tabulate their total win/loss value or score. The problem with this as can be noted in BoldFresh’s final standing placement as a top tiered player. One could easily argue that any victories against BF should have had the higher weighted value of “Y”.

Further, the reverse argument can be made for wins or losses against Zhukov if this format is utilized.

2. Tiers – who decides when one tier starts and another one ends? How many tiers should we have? What values do we use to assign to each tier? The placement of these tiers seems to be a critical component in the scoring system.
3. It was noted that a “weakness” in this system surrounds what should be done with players who only played 1 or 2 games. If we consider Anchovy, a strong player IMO, who didn’t participate much this year, or if we speculate that our old friend HOBO or Yoshi partook in a few games, but decided against completing more than 1 or 2, are we to weigh victories against these strong players as less valuable than those of players who had records such as 2-8, 5-15 etc….I would argue, not.
4. Even if we employ this new system or something similar, which by the way, was done back in 06-07, it appears that it will have a small impact on the overall and final rankings.

I certainly don’t want to seem indifferent to players who get “squeaked” out of the playoffs but in both systems three out of the four playoff participants are the same.

5. Considering the above, agreed to issues, players with 1-2 games played given a value, and the other “speculative” problems I note above, I would argue that adopting a new weighed system would be a bad idea.
6. Lastly, and in my opinion, this new scoring system goes against the KISS principle, of which, I am a strong advocate.
7. This is only my opinion on these “fun”, gaming matters. I will humbly abide and continue to participate by whatever ruling the powers that be or the masses desire.

[DarthMaximus]

@JWW:

A few potential concerns I have with the most recently suggested scoring system.

1. The newly proposed weighted system suggests utilizing the current winning percentage system as a means to assign point values to wins and losses.

I would argue, that if the current percentage based scoring system is deemed inequitable, then, any future system, using the percentage system as its foundation, would be similarly faulted.

This is actually a very strong point.

[666]

They were all strong points,  :-D I left a few “weak” ones off the table. If you want some of the weak ones let me know.

[666]

now get back to work, kicking my arse the our league game. I want to record the loss this year!! Stop stalling.

[Yoshi]

Kind of agree to JWW points, but also appreciate the detailed proposition and examples of everyone. I definitely do not appreciate the points system from revised ; too much value for the number of games played to my mind (but that’s just my opinion of course)

@JWW:

6. Lastly, and in my opinion, this new scoring system goes against the KISS principle, of which, I am a strong advocate.

What is this KISS system ?

[gamerman01]

KISS = Keep it simple, stupid

JWW, the distinction between Bold and OBG is the centerpiece of my argument.  Bold is a top tier player, and this is not seen in the straight won/loss record.  Yes, I know I used records to make tiers.  You guys - I took TWENTY MINUTES to devise the system and also calculate the rankings.  It’s the general idea that I’m putting out, a lot of the details are changeable, and it can be very simple, yes.

The straight won/loss record has a significant flaw - it punishes players who play tough opponents, and rewards players who bottom-feed.  A simple strength of schedule component (could be even simpler that what I threw out there today for thoughts) significantly increases the accuracy of the standings.

[Zhukov44]

I find JWW and DM’s arguments for KISS and simplicity quite convincing.  It seems to be true that both systems ‘would’ produce similar results.  Obviously they’ve put alot of thought into it so I will be fine with whatever they decide.

But……it seems to me that the best objection to JWW’s reasoning above is the ‘hypothetical’ one of a user who was bent on ‘gaming’ the system and only took games against noob opponents that he/she knew he could defeat.  It’s exactly this sort of user (if he/she appeared) that would produce a difference b/w a weighted and non-weighted system.  Since we are (mostly  :lol:) honorable in these parts, and the top players are willing to test themselves against other good players (to a limited extent anyway), there is essentially no difference between the weighted and non-weighted scores.  The potential for an exceptional case is what would make the various proposed extra rules and weighted scoring necessary.

Since we don’t seem to have much in the way of abuses, laisse-faire simplicity seems reasonable.  But if an abuser came along, I could see perceptions changing.

In other words, while previously the abuses have been indirect and more or less unconscious, if someone came along that was ‘directly’ and straightforwardly gaming the system, then Gamer’s argument that “The straight won/loss record has a significant flaw - it punishes players who play tough opponents, and rewards players who bottom-feed” might become more convincing than the KISS principle.

Good discussion all.