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
Yonni
Silver Donator
Silver Donator
Posts: 301
Joined: Thu Oct 19, 2017 6:55 pm
Karma: 220

Re: How often do games draw?

#21 Post by Yonni » Mon Dec 28, 2020 2:46 pm

Claesar wrote:
Mon Dec 28, 2020 7:59 am
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.
A 76% draw rate is too low?! That's about as high as I've seen. I've always thought I was too much of a carebear with a 67% draw rate but you're as kind as they come. The Master Gamer skewed the stats a little with his 50% win rate. I wonder how much of that speaks to differences on play here between his time and now.

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#22 Post by RoganJosh » Mon Dec 28, 2020 3:33 pm

jasnah wrote:
Mon Dec 28, 2020 8:26 am
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.
Nah, even if R only wants to secure Berlin you need England to backfill Kiel at some point, which means you need coordination, which means there's always the possibility of messing up. Since it's gunboat.

But you're arguing based on one example. Let me give you another example: in Valis' 21 game gunboat series, there where 3 solos thrown because of (expected) whittle attempts. And there weren't even that many whittle attempts on total, maybe 5 or 6, I don't really remember. That's what my opinion is based on.

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#23 Post by RoganJosh » Mon Dec 28, 2020 3:39 pm

Also, I think people get the numbers wrong. For every failed 4 > 3 whittle, you need three successful 4 > 3 whittles just to make up for the lost points. For every failed 5 > 4 whittle, you need four successful 5 > 4 whittles just to make up for the lost points. I think everyone should look through their own games and check whether they're actually gaining or loosing from whittling.

Edit: These are the games where you're one of the minor powers, the ones risking to lose the game. They are the ones deciding whether there will be a whittle or not. If you're the board to then you've got nothing to lose.

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#24 Post by RoganJosh » Mon Dec 28, 2020 3:47 pm

*board top

Claesar
Site Moderator
Site Moderator
Posts: 1923
Joined: Tue Oct 03, 2017 10:34 am
Karma: 1440

Re: How often do games draw?

#25 Post by Claesar » Mon Dec 28, 2020 8:46 pm

Yonni wrote:
Mon Dec 28, 2020 2:46 pm
A 76% draw rate is too low?! That's about as high as I've seen.
...
That stat is quite skewed too as most of my FP games were live. I expect there are fewer eliminations in those, due to time constraints.

In my definition, a carebear sacrifices solo chances to have a draw instead. This would be a carebear profile (anonimised, but it's someone on this site):
Won: 1 ( 1% )
Drawn: 47 ( 48% )
Survived: 22 ( 22% )
Defeated: 28 ( 29% )
Total (finished): 98

zultar
Co-Owner
Co-Owner
Posts: 79
Joined: Tue Mar 14, 2017 10:59 pm
Karma: 813

Re: How often do games draw?

#26 Post by zultar » Mon Dec 28, 2020 10:49 pm

Hey, I don't think it's nice to call me out like that. And then putting up my profile, damn. I am the carebear of webDip. They call me Mr. CareBear.
4

A_Tin_Can
Lifetime Site Contributor
Posts: 208
Joined: Fri Sep 29, 2017 9:18 pm
Karma: 356

Re: How often do games draw?

#27 Post by A_Tin_Can » Tue Dec 29, 2020 6:03 am

Zultar, I’m pretty sure that’s not what the mod team called you when I was on it...

Thanks teccles for doing the legwork on some updated stats! That’s very interesting. I’d be interested to see if three way draws have remained the most common draw type- at least in my games, larger draws now seem more common.

I do think that diplomacy play tends to go in fashion- I remember chatting with uclabb a few years ago about France doing well and Italy doing poorly (at the time)- we concluded that the webDip meta for Italy was to not go to Piedmont and that that probably had an impact (I still think that Italy does poorly with new players largely because there are disproportionately more articles with horrible advice for Italy than any other country).

I did look to see if I could make graphs of long term trends, but I couldn’t produce much that showed anything unexpected or especially interesting. I’d love it if someone was able to get a better “webdip meta” picture, though!

Fun aside: although (or because) the most common game result on webdip is/was a solo, the most common player result is a loss. This means that predicting “all players lose this game” outperformed GR (and all experimental Elo/GR variants I could think of) in terms of predicting individual player outcomes from a game.

Yonni
Silver Donator
Silver Donator
Posts: 301
Joined: Thu Oct 19, 2017 6:55 pm
Karma: 220

Re: How often do games draw?

#28 Post by Yonni » Tue Dec 29, 2020 12:41 pm

I’d love it if someone was able to get a better “webdip meta” picture, though!
I've got too much time on my hands these day. If you could get me access to a data dump, I'd love to try make something out of it.

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#29 Post by RoganJosh » Tue Dec 29, 2020 2:26 pm

FYI, I'm putting together one for gunboat games.

jasnah
Posts: 493
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 267

Re: How often do games draw?

#30 Post by jasnah » Tue Dec 29, 2020 2:44 pm

RoganJosh wrote:
Mon Dec 28, 2020 3:39 pm
Also, I think people get the numbers wrong. For every failed 4 > 3 whittle, you need three successful 4 > 3 whittles just to make up for the lost points. For every failed 5 > 4 whittle, you need four successful 5 > 4 whittles just to make up for the lost points. I think everyone should look through their own games and check whether they're actually gaining or loosing from whittling.

Edit: These are the games where you're one of the minor powers, the ones risking to lose the game. They are the ones deciding whether there will be a whittle or not. If you're the board to then you've got nothing to lose.
I would love if people did think this way as it would end whittling which we can probably agree is a good thing.

More power to those who threw solos in Valis21 as I've written above that I think people should be doing (especially you as Italy, which having read that game I'm 100% with your decision by the way). Ironically you wrote earlier that "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." which supports my point that the meta-wide whittling imperative created by DSS sucks regardless of whether or not a player in a given game can/will be whittled. Clearly whether or not the neighbours in these games would have chosen to whittle the player made no difference to their expected value in that game since the player, due to experiences with the toxic and asinine meta that follows directly from DSS, would have thrown either way!

The big problem for the playing experience comes about when players behave differently due to disagreement in playing philosophy: there's a category of players who believe in always utility maximising that particular game and will not throw pre-emptively even when their position appears a 90% chance of being whittled, since a 10% chance of getting points to them is still better than a 0% chance and outside game concerns do not factor into their decision. These players will continue support hold, support holding until they roll over and die a few turns later. The more they play, the higher the chance that anyone considering whether to whittle in a particular game has of getting away with it, and the more likely you are to be whittled. Then there's the other class of players like me (and I suppose you to a lesser extent) who will enthusiastically throw away our 10% chance of drawing if the other defenders move so much as one unnecessary unit toward us, in order to change the meta for the better. Maybe some people see this whittling minigame as a feature of Gunboat; I think it's a design flaw and we should do away with it entirely by playing a less rubbish scoring system.

As far as I can tell we're in full agreement that people should stop whittling, but I disagree with you on the grounds behind said conclusion. I hope I've adequately explained my thought process, which was developed by playing in and reading through hundreds of games. Of course yours is as well and so I respect that we have different board reads for example in the game I wrote earlier, which is no problem and normal for Diplomacy.

swordsman3003
Gold Donator
Gold Donator
Posts: 231
Joined: Tue Jan 02, 2018 2:51 pm
Location: Georgia, USA
Karma: 546
Contact:

Re: How often do games draw?

#31 Post by swordsman3003 » Tue Dec 29, 2020 8:59 pm

From my point of view, the incentive to whittle out players is a feature—not a flaw.

The term "draw-whittling" is loaded, in my opinion, because it implies an assumption that the players cooperating to eliminate a rival are seeking a draw per se. Plenty of players cooperate to eliminate a rival just to continue the match and see what happens next—or even as part of a strategy to get a solo win.

I view Diplomacy as a game about people, not a game about points and supply centers. "Who's next?" is such an interesting point of tension. Social deduction games like Mafia and Among Us contain a distillation of this one aspect of Diplomacy, and I think that's a major factor in why they are so fun to play.
3

AnimalsCS
Posts: 31
Joined: Sun Apr 26, 2020 7:59 pm
Karma: 49

Re: How often do games draw?

#32 Post by AnimalsCS » Tue Dec 29, 2020 9:38 pm

jasnah wrote:
Tue Dec 29, 2020 2:44 pm
I think it's a design flaw and we should do away with it entirely by playing a less rubbish scoring system.
I'd be interested in hearing your suggestions for alternate scoring systems that eliminate draw-whittling without changing the other incentives that DSS does a good job creating—encouraging solos and discouraging focus on how many centers you have in a draw.
2

jasnah
Posts: 493
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 267

Re: How often do games draw?

#33 Post by jasnah » Tue Dec 29, 2020 10:27 pm

swordsman3003 wrote:
Tue Dec 29, 2020 8:59 pm
From my point of view, the incentive to whittle out players is a feature—not a flaw.

The term "draw-whittling" is loaded, in my opinion, because it implies an assumption that the players cooperating to eliminate a rival are seeking a draw per se. Plenty of players cooperate to eliminate a rival just to continue the match and see what happens next—or even as part of a strategy to get a solo win.

I view Diplomacy as a game about people, not a game about points and supply centers. "Who's next?" is such an interesting point of tension. Social deduction games like Mafia and Among Us contain a distillation of this one aspect of Diplomacy, and I think that's a major factor in why they are so fun to play.
Yes, I realise you and some people enjoy this part of the game, however I and other people don't. DSS is fun for people in the former group and not in the latter.

I play Diplomacy to witness/attempt satisfying plays such as beautiful tactical combinations and for feeling of accomplishment in cooperating with other players (really), with the aim of growth, so eliminating players for the sake of it doesn't add much value to my experience.

Ari, I've written at length about your question elsewhere, but I'll return to it here when I have time.

Aristocrat
Posts: 12
Joined: Sun Jul 19, 2020 6:09 pm
Karma: 13

Re: How often do games draw?

#34 Post by Aristocrat » Tue Dec 29, 2020 10:45 pm

swordsman3003 wrote:
Tue Dec 29, 2020 8:59 pm
From my point of view, the incentive to whittle out players is a feature—not a flaw.
I agree with this (though I almost exclusively play press). My stats rival the carebears' posted upthread, but if anything I think we need more whittling, not less. The threat of being whittled down the road, and trying to avoid it, introduces another strategic dimension to the game even if a solo is out of reach. Failure to whittle leads to the current situation where people leave games half played and it encourages a lazy "it's 1904, let's get the 4 strongest countries and just draw this" mentality. I'm happy to take advantage of the laziness meta when the opportunity presents itself (often) but it doesn't mean it leads to fun games.

AnimalsCS
Posts: 31
Joined: Sun Apr 26, 2020 7:59 pm
Karma: 49

Re: How often do games draw?

#35 Post by AnimalsCS » Wed Dec 30, 2020 4:33 am

jasnah wrote:
Tue Dec 29, 2020 10:27 pm
Ari, I've written at length about your question elsewhere, but I'll return to it here when I have time.
Discord me a link? I'd be interested to read your thoughts on the matter.

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#36 Post by RoganJosh » Wed Dec 30, 2020 5:22 am

swordsman3003 wrote:
Tue Dec 29, 2020 8:59 pm
Plenty of players cooperate to eliminate a rival just to continue the match and see what happens next—or even as part of a strategy to get a solo win.
You're taking the viewpoint of the solo contender. And for the solo contender it's of course a "feature": they get a second shot at the solo with no attached risk.

But out of curiosity, let me ask: in these games where you're the solo contender, you've reached a stalemate line, and now you back off to give yourself a second shot at the solo. How often do you solo? And how often do you have to settle with a draw (whatever size)?

RoganJosh
Silver Donator
Silver Donator
Posts: 536
Joined: Sun Dec 31, 2017 1:02 am
Location: Stockholm
Karma: 434
Contact:

Re: How often do games draw?

#37 Post by RoganJosh » Wed Dec 30, 2020 6:32 am

@jasnah

We're definitely in agreement on most things. But I do think your point is somewhat strange. My guess is that you have a FTF background, and that you're conflating the FTF DSS meta with the online DSS meta. I might be completely wrong, in which case you'll have to forgive me.
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.
My impression is that you've realized that whittling isn't actually such a big problem here - at least not compared to FTF - but you can't explain why. So you ascribe it to an 'honorable player culture', as if people here don't understand their own scoring system. (And, as if people here are honorable...)

The solo contender should of course always invite the whittle - never deprive your opponents a chance to make a mistake. My point, as you already know, is that it's rational for the minor powers to decline the whittle. There are some obvious reasons:

1. Too small reward, too big risk. We've been through this one already.

2. (Meta) Also small powers are needed to stop solos, but if you always whittle them, then they won't help out. You basically mentioned this one yourself.

As for players who think that they should optimize each and every game, that's not an effect of DSS. And if someone is so stupid as not to put on a fight when they're being whittled, well, then they're not even playing to optimize that one game.

Assuming that you're from a FTF background, let me point out two other reasons for why declining the whittle is more rational online than FTF.

3. There's an abundance of games. Sure, you can put a lot of effort into bumping .20 up to .25. Or you can just take the .20 and start a new game. What'll give you the best return per effort? Compared to risk?

4. You can't sustain an inflated ranking. FTF, the ranking list is always finalized in a not-too-distant future. In an ELO-based and constantly ongoing ranking system, an inflated score will be lost as you play more games. That is, optimizing the return from one game is less important than improving your overall skill. (Unless you do it like that vDip guy, who simply quit after getting an inflated score that put him top of the rankings.)
1

jasnah
Posts: 493
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 267

Re: How often do games draw?

#38 Post by jasnah » Wed Dec 30, 2020 7:45 am

I've never played ftf. Vftf yes, but I wouldn't expect the dynamics to play out in the same way. My personal dislike of DSS comes exclusively from playing and observing it on this site. I have not played DSS anywhere else, although I share my reasoning with much of the ftf community where DSS is now obsolete.

Yes, the solo contender has only to gain by inviting the whittle, similarly the solo contender has only to gain by refusing to draw into perpetuity and waiting for misorder or attrition. Fortunately the site has rules against this, the point being that just because something is allowed according to the official rules of Diplomacy and personally beneficial to one of the players doing it, doesn't make it desirable for the game overall.

You've given your take on why whittling is actually not a rational thing to do. I don't think this view is correct, nor do I think it's shared by the majority of players. Whittling is often no more effort than entering easy orders for a couple more years and there are players on this site who think that the only way to maximize points and rating (not just in one game but overall) is by leaving no points on the table, ie by sucking the last drop of value out of superfluous powers' survival bonus. It's also minimal risk when the solo threat properly facilitates it by moving back, making it trivial matter for the minor powers to do it. I can see that we still disagree on the risk-reward premise you've cited and maybe I've just been playing with people who are not as vindictive as the people you've been playing with. Either way the outcome of us having "been through this one already" is that we still disagree on it.

jasnah
Posts: 493
Joined: Tue Jun 02, 2020 8:13 pm
Karma: 267

Re: How often do games draw?

#39 Post by jasnah » Wed Dec 30, 2020 8:02 am

AnimalsCS wrote:
Wed Dec 30, 2020 4:33 am
jasnah wrote:
Tue Dec 29, 2020 10:27 pm
Ari, I've written at length about your question elsewhere, but I'll return to it here when I have time.
Discord me a link? I'd be interested to read your thoughts on the matter.
Scattered across a number of Discord servers but one of them is the WebDip discord.

Anyway to answer your question, I'm interested to know why you think DSS encourages solos, because as far as I can figure DSS is a rather poor system for encouraging people to go after solos - Carnage and to lesser extent the Tribute family due to board topper tribute incentives are both better. In DSS meanwhile the reward for making early decisions that lock you into a high chance of a 3wd (over a 100% ROI with webDip points) but a minimal chance of a solo is sufficient that many players would and do take it over the somewhat higher chance of a solo with a significant chance of elimination or losing to someone else's solo. While this is debatable how much is a feature of DSS and how much is due to people's inherent risk aversion, DSS does a pretty poor job of punishing the regularly lambasted draw-mongering style of play.

I've been persuaded in favour of WTA PPSC (solo = 34 points and others zero, else points = no of dots) which has the following benefits:

1. Elementary to calculate score under all circumstances
2. Rewards non-solos at no more than 1/2 the points of solos, like DSS, and unlike the Tribute family
3. Meaningful points for small powers to fight to defend what they have, unlike squares
4. Zero incentive to whittle for the sake of it

Yes, this gives larger reward to those who have more centres in a draw. I have no problem with this. I consider someone who ends in the draw with 17 centres to have had a more successful game than someone who ends in the draw with 2 centres, which funnily enough is the same conclusion that DSS comes to, because the 2 centre player is almost certainly not making it.

If this site switched to WTA PPSC as the default what would happen is that good players would still be good and understand position not start behaving like headless chickens, and bad players would still be bad, as they tend to be.
1

Claesar
Site Moderator
Site Moderator
Posts: 1923
Joined: Tue Oct 03, 2017 10:34 am
Karma: 1440

Re: How often do games draw?

#40 Post by Claesar » Wed Dec 30, 2020 8:30 am

I've found that Carnage scoring prevents solos in practice. So many early draws because people are content with their current rank..

Post Reply

Who is online

Users browsing this forum: No registered users and 7 guests