How often do games draw?

New players can go here for helpful advice and to sign up for our mentor program, or if you're a veteran help answer questions.
Forum rules
This is an area for new members or members looking for help with the site or Diplomacy. Off topic threads and replies will be moved to the appropriate category.
Message
Author
Hekeho13
Posts: 1
Joined: Thu Dec 24, 2020 3:43 am
Karma: 2
Contact:

How often do games draw?

#1 Post by Hekeho13 » Thu Dec 24, 2020 3:46 am

Hello!

I'm playing with a group of friends, and somebody mentioned that the point of Diplomacy is to always go for a win, or in rare cases, a double-win because of a stalemate. They made it sound like it's very hard to get draws and it's only ever really one or two people.

However, I'm looking at a few tournament games and games on this forum and I see a lot of 4-way and 5-way draws. Why is that? Is this because there's an alliance and basically, the alliance won't go against each other and if two "alliances" are equal power, they draw?

My friend made it sound like it's inappropriate to ever make an alliance that goes all the way to a draw, and that in an alliance if one side is stronger than another, you should eat up that weaker side and become huge.

Can someone help me understand the general rule of thumb when it comes to draw? Currently, I've been told Diplomacy should be a single winner or two-winner if tie, and never ever more.
2

AnimalsCS
Posts: 32
Joined: Sun Apr 26, 2020 7:59 pm
Karma: 51
Contact:

Re: How often do games draw?

#2 Post by AnimalsCS » Thu Dec 24, 2020 4:13 am

Your friend is definitely incorrect about solos and two-way draws being the normal results for a game of standard diplomacy. Three way draws are most common for a variety of reasons, even when all players desire a solo.

Here are a couple excellent articles that might help explain why:
http://uk.diplom.org/pouch/Zine/S2001M/ ... Draws.html
https://brotherbored.com/gunboat-diplom ... ate-lines/

Finally, on the point of alliances, context is everything. Assuming two players in an alliance are both playing for a solo, it is nevertheless entirely possible that neither ends up finding an appropriate time to stab the other for the win and the alliance lasts for the entire game. It is also possible for an alliance to last for a few turns, or for a weaker power to actively try to ally with a stronger power, despite the inherent risks.

There is a lot of nuance in this game and a lot to learn. I've found BrotherBored (website) and FloridaMan (youtube channel) as particularly helpful in my improvement as a player but there are many other resources out there too.
10

gimix
Site Moderator
Site Moderator
Posts: 349
Joined: Tue Nov 13, 2018 8:56 am
Location: My Mountains, the highest in Europe
Karma: 343
Contact:

Re: How often do games draw?

#3 Post by gimix » Thu Dec 24, 2020 6:47 am

Just to be clear, there is no thing like a double-win, so, yes, your friend is incorrect. There are only three possible results: victory, which means solo win, draw, be it 2, 3, ..., 7-way, and loss, which may mean that you've been eliminated or that you survived to a solo.

Solo win is the target - then again it is a difficult target, and if everybody plays well, it is very rare.
Draws are common: 2-way are infrequent because of the precariousness of the 17-17 equilibrium; 3 or 4-way are the norm; "majority wins", ie draws with more than 4 players included are legitimate but slightly frowned upon as being inelegant by many players
Losses are losses: even being a "strong second" in someone else's win is not an accomplishment - someone would even say it proves that you didn't play well since you failed to stop the solo!
3

President Eden
Posts: 6907
Joined: Fri Oct 20, 2017 2:11 pm
Location: possibly Britain
Karma: 9609
Contact:

Re: How often do games draw?

#4 Post by President Eden » Thu Dec 24, 2020 7:24 am

Hekeho13 wrote:
Thu Dec 24, 2020 3:46 am
Hello!

I'm playing with a group of friends, and somebody mentioned that the point of Diplomacy is to always go for a win, or in rare cases, a double-win because of a stalemate. They made it sound like it's very hard to get draws and it's only ever really one or two people.

However, I'm looking at a few tournament games and games on this forum and I see a lot of 4-way and 5-way draws. Why is that? Is this because there's an alliance and basically, the alliance won't go against each other and if two "alliances" are equal power, they draw?

My friend made it sound like it's inappropriate to ever make an alliance that goes all the way to a draw, and that in an alliance if one side is stronger than another, you should eat up that weaker side and become huge.

Can someone help me understand the general rule of thumb when it comes to draw? Currently, I've been told Diplomacy should be a single winner or two-winner if tie, and never ever more.
The most accurate answer requires analyzing the incentives behind the games played. For instance, webDiplomacy's most traditional scoring system is "Draw Size Scoring," which means every surviving player in a draw gets rewarded, but a winner gets the whole pot. For a while we had a second system called "Points Per Supply Center" (which has been discontinued), which means players are rewarded per supply center, regardless of who wins or loses. PPSC games had a lot of "strong seconds" where you could easily win from 14-15 if a second player had 10+ centers, because the second player has very little incentive to fight for the draw. DSS games are the tournament norm, and draws are very common because the players all know how to fight solos and get rewarded for doing it. This is oversimplification cuz I'm up late but it's getting the gist.

In general, the games tend toward draws because highly defensible stalemate lines exist. Especially in more experienced games, everyone KNOWS how to form the lines and has a good sense of anticipating when a solo is coming well in advance. So unless someone wants to throw - which happens more often than you'd think - then it's almost impossible to win.
4

DrugTito
Posts: 21
Joined: Tue Dec 03, 2019 12:14 am
Karma: 16
Contact:

Re: How often do games draw?

#5 Post by DrugTito » Fri Dec 25, 2020 12:46 am

Seems like you have the perfect chance to pretend you bought what your friend was selling and stab him in a right moment for the actual win.
3

A_Tin_Can
Lifetime Site Contributor
Posts: 283
Joined: Fri Sep 29, 2017 9:18 pm
Karma: 451
Contact:

Re: How often do games draw?

#6 Post by A_Tin_Can » Sat Dec 26, 2020 11:55 am

AnimalsCS wrote:
Thu Dec 24, 2020 4:13 am
Three way draws are most common for a variety of reasons, even when all players desire a solo.
This is not correct. On webdiplomacy, around 60% of games end in a solo. I can't check the stats today, but in c. 2015 that was the case (even if you remove games with NMRs). I don't think there would be any reason for it to be different today.

On vDiplomacy, around half the games end in a solo: https://vdiplomacy.net/stats.php?variantID=1

It is true that many tournament or "high level" games end in draws. I suspect this is because of fear or grandstanding rather than because good players always force draws.

A_Tin_Can
Lifetime Site Contributor
Posts: 283
Joined: Fri Sep 29, 2017 9:18 pm
Karma: 451
Contact:

Re: How often do games draw?

#7 Post by A_Tin_Can » Sat Dec 26, 2020 12:01 pm

Before people point it out, I believe these stats also held true if you only looked at WTA (what we now call DSS) games. This isn't just because of PPSC (Although President Eden's (hi Eden!) point is strong).

teccles
Posts: 237
Joined: Mon Mar 25, 2019 8:10 pm
Karma: 285
Contact:

Re: How often do games draw?

#8 Post by teccles » Sat Dec 26, 2020 1:00 pm

I was curious about this, and it looks like it is no longer true that most games are solos.

I did searches like this one to check how many wins and draws there are in DSS and PPSC, both in general and with bet size 101+ (which is one easy way to get some kind of quality bar - though I don't know how useful this is, because there aren't that many games like this).

In PPSC, 63.5% of games were solos (17195 wins, 9889 draws), including 60.6% of 101+ point games (246, 160).

In DSS, 42.2% of games were solos (5370, 7355), and only 20.3% of 101+ point games (57, 224).
5

jasnah
Posts: 507
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 308
Contact:

Re: How often do games draw?

#9 Post by jasnah » Sat Dec 26, 2020 2:08 pm

I can't speak for Press but I play a lot of Gunboat and would indeed expect well played WTA (both DSS and SoS) to almost always end in draws. This is because both of these scoring systems, SoS to a even greater extent than DSS, reward playing in a style that leads to a likely solid position in a draw with a slight chance of solo rather than going for the race to 18 line-busting alliances that almost guarantee a solo endgame. Some people, in my experience a minority, have higher risk tolerance than me and would go for those alliances anyway on the gamble that they outplay their partner at the end, which is a valid choice.

I play mostly private games with OK to excellent skill level and the only solos I experience are those thrown by players who were going to be cut out of the draw or by someone blundering and playing the patsy in a lopsided alliance.
1

teccles
Posts: 237
Joined: Mon Mar 25, 2019 8:10 pm
Karma: 285
Contact:

Re: How often do games draw?

#10 Post by teccles » Sat Dec 26, 2020 2:19 pm

Yes - my statistics above were for press, but the same trends hold for gunboat:

PPSC (unfiltered): 69% solo
DSS (unfiltered): 44%
PPSC (101+): 75% - but extremely small sample of 67 games.
DSS (101+): 32%.

One additional note: in press, the PPSC solo rate was initially much higher; splitting the games into two equal sections by time, the solo rates are around 70% and 50%. So while the scoring system explanation seems very reasonable (the scoring system that encourages throwing games should indeed lead to solos), clearly some other things changed over time.
3

jasnah
Posts: 507
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 308
Contact:

Re: How often do games draw?

#11 Post by jasnah » Sat Dec 26, 2020 2:35 pm

I also think that DSS should encourage more throwing of solos than it does, bringing the solo rate to higher than squares, but this is a normative value judgment and people clearly don't agree with me as squares GB has a higher 50.1% solo rate across all games (too few games to look at 101+ pot).

AnimalsCS
Posts: 32
Joined: Sun Apr 26, 2020 7:59 pm
Karma: 51
Contact:

Re: How often do games draw?

#12 Post by AnimalsCS » Sat Dec 26, 2020 5:00 pm

teccles wrote:
Sat Dec 26, 2020 1:00 pm
I was curious about this, and it looks like it is no longer true that most games are solos.
Thanks teccles for checking the data on this, it's very cool that President Eden's explanation for why there were more solos in PPSC really reveals itself in the actual numbers. I've never played under that scoring system, so I didn't even think about discussing it in my initial response to OP.
jasnah wrote:
Sat Dec 26, 2020 2:35 pm
I also think that DSS should encourage more throwing of solos than it does, bringing the solo rate to higher than squares, but this is a normative value judgment and people clearly don't agree with me as squares GB has a higher 50.1% solo rate across all games (too few games to look at 101+ pot).
My guess is that there are way more cases where someone is on 1-3 centers and will get basically nothing from SoS but is in a position to throw the solo than there are cases where players try to eliminate someone from the draw in DSS in a situation where that player is in the position to throw a solo. For example: a French player who only has Por and Spa is unlikely to throw a solo in DSS since they have a key position on the stalemate line and will survive to a draw, whereas in SoS this same player gets almost no points and might decide to throw to the Italian player to prevent their English rival from winning/drawing.
1

jasnah
Posts: 507
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 308
Contact:

Re: How often do games draw?

#13 Post by jasnah » Sun Dec 27, 2020 8:28 am

AnimalsCS wrote:
Sat Dec 26, 2020 5:00 pm
teccles wrote:
Sat Dec 26, 2020 1:00 pm
I was curious about this, and it looks like it is no longer true that most games are solos.
Thanks teccles for checking the data on this, it's very cool that President Eden's explanation for why there were more solos in PPSC really reveals itself in the actual numbers. I've never played under that scoring system, so I didn't even think about discussing it in my initial response to OP.
jasnah wrote:
Sat Dec 26, 2020 2:35 pm
I also think that DSS should encourage more throwing of solos than it does, bringing the solo rate to higher than squares, but this is a normative value judgment and people clearly don't agree with me as squares GB has a higher 50.1% solo rate across all games (too few games to look at 101+ pot).
My guess is that there are way more cases where someone is on 1-3 centers and will get basically nothing from SoS but is in a position to throw the solo than there are cases where players try to eliminate someone from the draw in DSS in a situation where that player is in the position to throw a solo. For example: a French player who only has Por and Spa is unlikely to throw a solo in DSS since they have a key position on the stalemate line and will survive to a draw, whereas in SoS this same player gets almost no points and might decide to throw to the Italian player to prevent their English rival from winning/drawing.
You rarely survive a draw in just Por and Spa. It relies on the benevolence of opponents who care more about fair play than maximizing score. I’m very happy for every person who’s actually able to survive the draw in just Iberia and for the opponents who didn’t whittle them, but that’s strictly due to an honourable player culture and persists despite, not because of, DSS.
2

swordsman3003
Gold Donator
Gold Donator
Posts: 264
Joined: Tue Jan 02, 2018 2:51 pm
Location: Texas
Karma: 607
Contact:

Re: How often do games draw?

#14 Post by swordsman3003 » Sun Dec 27, 2020 8:13 pm

Taking a completely unscientific, pure-intuition average of all scoring system, all skill levels, gunboat and press, tournament vs. private vs. pickup games...

I expect about half of the matches I play in to end in draws, and about half of them to end in solos.

Keeping in mind that there are 7 players on the classic board, I consider a 7%-10% win rate to be a sign of a decent player, a 10%-13% win rate to be a badge of a strong player, and a 14%+ win rate to be evidence of phenomenal ability. Again, I am stating my subjective intuition; this view is not supported by data.

(By the way thanks for the compliments; I am BrotherBored)
1

Yonni
Silver Donator
Silver Donator
Posts: 408
Joined: Thu Oct 19, 2017 6:55 pm
Karma: 368
Contact:

Re: How often do games draw?

#15 Post by Yonni » Mon Dec 28, 2020 2:09 am

Looking at the top 10 player by ghost rating, their average draw rate is 47.5%. That means the draw rate in their games is at least that (but not much higher because they don't have a ton of games that they get eliminated from that end in draws).

Their average win rate is also a fairly remarkable 28%.

RoganJosh
Silver Donator
Silver Donator
Posts: 556
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 464
Contact:

Re: How often do games draw?

#16 Post by RoganJosh » Mon Dec 28, 2020 2:17 am

jasnah wrote:
Sun Dec 27, 2020 8:28 am
I’m very happy for every person who’s actually able to survive the draw in just Iberia and for the opponents who didn’t whittle them, but that’s strictly due to an honourable player culture and persists despite, not because of, DSS.
Are we talking gunboat or full press? With or without a time limits?

The most common situation in DSS is a 4WD that could be whittled down to a 3WD. Now, every whittle comes with a risk of a successfully thrown solo. And the math is pretty simple: if Mrs. solo contender only needs to get one 50/50 guess right to win, then the smaller powers should not try the whittle. If she needs to get two guesses right, then it's break even. Makes more sense to not whittle in that case. But if she needs three guesses or more, then the minor powers should whittle.

If it's a full press game, then the minor powers can usually coordinate and give conditions to the solo contender, to secure the whittle. If it's a gunboat game, then they can't. And it usually comes down to one or two guesses. Hence my motto: Never whittle in gunboat.

jasnah
Posts: 507
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 308
Contact:

Re: How often do games draw?

#17 Post by jasnah » Mon Dec 28, 2020 4:42 am

RoganJosh wrote:
Mon Dec 28, 2020 2:17 am
jasnah wrote:
Sun Dec 27, 2020 8:28 am
I’m very happy for every person who’s actually able to survive the draw in just Iberia and for the opponents who didn’t whittle them, but that’s strictly due to an honourable player culture and persists despite, not because of, DSS.
Are we talking gunboat or full press? With or without a time limits?

The most common situation in DSS is a 4WD that could be whittled down to a 3WD. Now, every whittle comes with a risk of a successfully thrown solo. And the math is pretty simple: if Mrs. solo contender only needs to get one 50/50 guess right to win, then the smaller powers should not try the whittle. If she needs to get two guesses right, then it's break even. Makes more sense to not whittle in that case. But if she needs three guesses or more, then the minor powers should whittle.

If it's a full press game, then the minor powers can usually coordinate and give conditions to the solo contender, to secure the whittle. If it's a gunboat game, then they can't. And it usually comes down to one or two guesses. Hence my motto: Never whittle in gunboat.
Or, you know, she does this https://webdiplomacy.net/board.php?gameID=328454 with barely any cost or risk to herself so that there's no chance of her ever making those guesses. This is the most ridiculous I've ever seen it, look at how far the Turkish units receded. I understand that this is the way for the Turkish player to points maximize but I found it totally distasteful.

I don't whittle anymore not because I'm under any delusion that it's not possible, when there is the desire it's almost always tactically possible, but because I don't want to.

RoganJosh
Silver Donator
Silver Donator
Posts: 556
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 464
Contact:

Re: How often do games draw?

#18 Post by RoganJosh » Mon Dec 28, 2020 5:25 am

I'm not sure why you are focusing on Turkey. In that game, it is Russia (you!) that decides whether there will be a whittle or not. And that's a 5WD, so you need to be 80% sure of succeeding in the whittle to try. And you'll have to assume that T will attack you as soon as you attack G. Doesn't look like you can do it without England, so there's always the risk of miscommunication. More than 80% of succeeding? Probably, since E moved to Burgundy. 100%? No way.

Maybe I don't play enough games nowadays, but I've seen more whittles fail than succeed. Even worse, I've seen whittlers who were so sure they would be whittled that they threw games where other players weren't even trying to whittle them.

Claesar
Posts: 1965
Joined: Tue Oct 03, 2017 10:34 am
Karma: 1490
Contact:

Re: How often do games draw?

#19 Post by Claesar » Mon Dec 28, 2020 7:59 am

Yonni wrote:
Mon Dec 28, 2020 2:09 am
Looking at the top 10 player by ghost rating, their average draw rate is 47.5%. That means the draw rate in their games is at least that (but not much higher because they don't have a ton of games that they get eliminated from that end in draws).

Their average win rate is also a fairly remarkable 28%.
That is most interesting! I thought my draw rate was too low, but it appears my solo win rate is actually further off the mark.

jasnah
Posts: 507
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 308
Contact:

Re: How often do games draw?

#20 Post by jasnah » Mon Dec 28, 2020 8:26 am

RoganJosh wrote:
Mon Dec 28, 2020 5:25 am
I'm not sure why you are focusing on Turkey. In that game, it is Russia (you!) that decides whether there will be a whittle or not. And that's a 5WD, so you need to be 80% sure of succeeding in the whittle to try. And you'll have to assume that T will attack you as soon as you attack G. Doesn't look like you can do it without England, so there's always the risk of miscommunication. More than 80% of succeeding? Probably, since E moved to Burgundy. 100%? No way.

Maybe I don't play enough games nowadays, but I've seen more whittles fail than succeed. Even worse, I've seen whittlers who were so sure they would be whittled that they threw games where other players weren't even trying to whittle them.
There’s no way for Germany to throw two (more after Russia gets back Moscow) dots to Turkey after he moves back that far. Turkey is never able to pick up Stp and he can’t get more than one of Munich or Berlin, even one is a stretch requiring a brilliant German move and coordination. Your desire to rationalise your ethical belief against whittling is influencing your assessment of board state.

Post Reply

Who is online

Users browsing this forum: No registered users and 43 guests