webDiplomacy’s Scoring Systems

webDiplomacy is one of the top Diplomacy sites in terms of numbers. It has a lively forum – much improved over recent years, since it adopted the phpBB forum structure. It has a lot going for it, including a membership that doesn’t seem to put up with the abuse that was pretty much unchecked on the forum, now seemingly a thing of the past. Having said that, it isn’t the site I prefer to play on. There may be things going on that I don’t know about. Frankly, I’m not interested in knowing about it if this is the case!

No site is perfect. While other people can tell you what is great about webDip, I’m going to put a few posts together about what is disappointing about it. Again, because it isn’t my site of preference, it may be that there’s more that I don’t like.

[By the way, before any webDippers get after me for taking this approach, I’m going to do the same for Playdiplomacy, too. In fact, I have more to criticise about Playdip, probably because I play there more and prefer it!]

The first disappointing aspect of webDiplomacy is the scoring system… well, actually systems because there are two of them (three if count the semi-official True Skill-style Ghost Ratings).

This double ratings system is, in itself, a problem. When you create a game, you have to choose the type of scoring system the game will use. These are the “Draw Sized Scoring” system (which was previously called “Winner Takes All” until it was re-branded along with a fair amount on webDip) and the Sum of Squares scoring system.

The basic scoring system

webDip has a unique philosophy behind their systems, which is a wager system. You have a base of 100 points. You have to wager a number of these points to enter a game. The smallest number of points a player can wager is 5; this means that you can enter a maximum of 20 games at any one time.

If you drop below 100 points, though, you’ll be topped-up and return to 100. How this works if you’re in 20 games and one ends, with you losing, I don’t know. Do you return to 100 points, so adding another potential 20 games? This is how the system is described. I wonder, though, if in reality your points aren’t simply reimbursed.

The idea of this is that you can always play in scored games. Personally I don’t think this is needed. Why would anyone want to play in 20 games at once? Even with the Gunboat variant, which doesn’t require any communication, to do justice to the games there is a limit.

For me, they really ought to make the minimum number of points a player can commit 20. This limits the number of games to a much more manageable 5 consecutive games. This might help prevent drop outs.

I’d also not want to reimburse points if a player doesn’t have any points left to wager. Use up your points – be reduced to 0 – and you have to play unranked games to recuperate your points. The reason points are recuperated is that you require points to play unranked games.

Why two systems?

Honestly, I don’t know. it looks like they can’t make their minds up, frankly. Either that or the site is more interested in placating members who don’t like DBS systems, or who dislike SCS systems.

The even more confusing aspect is that the points from these two different styles of systems are combined to give a single points score. Why they aren’t separated and two ranking systems used, I’ve no idea. The two systems reward very different types of play.

webDip acknowledges the differences in play:

Draw-Size Scoring encourages players to play for a win, but if they cannot win, it encourages players to “narrow” down a draw by eliminating smaller powers in order to get a better result. With fewer players remaining in the draw, each remaining player receives more points of the pot.

Points on webDiplomacy

Narrowing down the draw is a polite way of saying that players need to eliminate others from the game to maximise the result. It encourages 2-way draws – which is an outcome that really ought to be extremely rare in Dip games. DBS scoring can mean games drag on, while one or more attempt to eliminate the opposition. This is an issue when the game clearly isn’t going to be resolved.

Sum-of-Squares encourages players to play to win, but if they cannot it does not drag on games longer just to get another elimination because the number of players in the draw is less critical than the number of supply centers each player in the draw owns. Sum-of-Squares is often used in face-to-face Diplomacy tournaments where games may be time sensitive.

Points on webDiplomacy

SoS scoring is about grabbing supply centres. You will need a strong alliance to do well, of course. At the end of the game, though, stealing an SC from your ally might put you in a better position. It can be argued that this makes for more exciting play, with a lot of fluidity of alliances, and it probably does. What it doesn’t do, however, is mean games aren’t “dragged on” necessarily.

For instance, let’s say that you and your ally are doing well. You’re on 12 SCs, your ally on 11. You’re unlikely to win the game: stabbing your ally to go for a win is dangerous, and the opposition is strongly united. Why not end the game now? This is more likely under a DBS system, frankly.

When using SoS, the temptation is to assume the game will end in a draw, and play to grab that extra SC. Instead of playing to maintain the alliance, then, you’re more likely to play to break it. Establishing a 2 SC lead over the player in second will increase your score.

What this may lead to is a prolonging of the game. Your erstwhile ally, realising that they’ve just lost significant ground to you, might well switch sides and begin working with the opposition. If they’re vindictive, this might well extend the game until you’re eliminated!

Saying SoS will prevent games from being “dragged on” isn’t accurate. In fact, it may well encourage more cautious play than is assumed, because it means that players in alliance play more cautiously simply to prevent a potential snatch-and-grab SC raid.

SoS was introduced to replace an alternative system that webDip called Points Per Supply Centre (PPSC):

webDiplomacy used to support a scoring system called Points-Per-Supply-Center, which was a supply-center-based scoring system that distributed the pot based on a ratio of the number of supply centers each player owned. This scoring system was discontinued because of the prevalence of “strong seconds,” which is when one player wins because another player is promised more supply centers, and thus more points, if they help the other player win. Since this goes against the intent of the game rules this scoring system was removed.

Points on webDiplomacy

SoS is certainly an improvement on this system… but it isn’t the answer to it, frankly. SoS is also an SCS system, although modified to make it more complex. I’m not sure why SoS removes the problem webDip correctly found using PPSC.

I think it should also be pointed out that the idea webDip calls the “Strong Second” idea isn’t accurate. Under Strong Second systems, when a game ends in a solo, credit is given to players who ended the game on most SCs behind the soloist.

The problem with SoS scoring here

There is an additional problem with these systems: they are scoring systems intended for a short series of games rather than an on-going series. They are tournament systems, not ratings systems.

In a tournament, which might cover 4 or 5 games, SoS has the advantage of a lot of other scoring systems in that it is based on the final outcome of the game and it effectively differentiates between games. A player in one game, ending in a 4-way draw, with 13 SCs at the end, is likely to score differently then another player in another game with the same outcome. This is because SoS scoring depends on comparing the number of SCs held by surviving powers.

However, when this system is used to score a long series of games, and especially an on-going series, it doesn’t work. SoS comes into its own when games end in draws, which is the vast majority of tournament games. When enough games end in solos, the differentiation becomes less useful.

The other aspect of SoS scoring is that it changes the way the game is played. With this system, assuming the game will end in a draw, it becomes more about increasing SC count. SoS – and any other SC count system of scoring – ignores the rule that, if a game ends before someone wins, all players share equally in the draw.

Arguably this is acceptable in a tournament; there’s no defence for this in an on-going series of games. If the game ends in a draw, it doesn’t matter how many SCs a player has – nobody has reached the objective of owning 18 SCs.

Hybrids rarely work

The big crime here is using two different scoring systems, which encourage two different approaches to the game, and mashing them together as if they are complimentary.

If you want you can play on webDip by entering games with just one scoring system – you could play only DSS games or only SoS games. However, you can’t escape the fact that the site will rank you based on a combination of games using both.

Of course, you could also only play in unscored games but then, as is common to any game site, you’re likely to play lower quality games because most players want to see how they compare to others.

The answer, of course, is to play each game as if you weren’t interested in scoring and to just play as if it was really a stand alone game. After all, that’s how it was designed. That’s how I tend to play, unless I’ve entered a tournament.

This doesn’t take away from the fact that webDip forces players who want to play in scored games deal with two opposing scoring systems resulting in two opposing play styles. To me – obviously – this is ridiculous; I’d love to hear why this decision was made.