@ carpenter - my apologies, I was not intending to be suggestive in any way.

And RR calculates this way:
Reliability rating is 100 minus phases missed / phases played * 200, not to be lower than 0
Example: If a user misses 5% of their games, rating would be 90, 15% would be 70, etc.

As usual, Oli's got the clearest vision of things and I agree with the reasons that brought him to the current system.

'Though I have to say that I'm really thinking of leaving the game I told (the one with a second turn NMR) because it's totally screwed and Cat & Fox are not voting Cancel. So, since both ways someone would be unhappy...

But the important thing Oli said is this:
>I tried to think about bigger penalties for the country going CD, but usually
>its a new player that missed to check the site regularly. If I give him too
>many missed phases he won't be able to enter new games with a bad RR
>and I have to give him points back. And if it's a long time player some more
>missed phases don't make a great difference.

Guys, am I the only one who sees that a PROPORTIONAL PENALTY is a solution?
It would have always the same impact on the RR!

Don't forget we want to punish serial/compulsive criminals, not those unlucky players who had, once ot twice in a year, some real life issue. The idea would be this.

- NMR is when you miss orders one turn (not the first or second of the game)
- CD is when you miss orders in the first or second turn of the game OR the second time in-a-row in the same game.

Now, say we take out 1% of phases for NMR and 3% for CD (rounded up). And say we have 3 players who have never missed a turn:

A - RR=100% (missed 0 of 24 phases)
B - RR=100% (missed 0 of 229 phases)
C - RR=100% (missed 0 of 1127 phases)

Now they NMR, so:

A) 1%x24 = 0.24 Round Up = 1. So "(missed 1 of 24 phases)", then the current formula would give RR = 92%.
The huge rounding up may make it look a heavy penalty, but since this player just joined the site and played so little phases, he will recover his RR quickly.

B) 1%x229 = 2.29 Round Up = 3. So "(missed 3 of 229 phases)", then the current formula would give RR = 97%

C) 1%x1127 = 11.27 RU = 12. So "(missed 12 of 1127 phases)", then the current formula would give RR = 98%

In the latter cases , due to the high number of phases played, RR will be slower to recover and that's why the penalty is lighter. But not so light to be meaningless!

Let's see CDs:

A) 3%x24 = .72 Round Up = 1. So "(missed 1 of 24 phases)", then the current formula would give RR = 92%.
The same of NMR? Yes! He's a noob, let's give him time to understand how this site works!

B) 3%x229 = 6.87 Round Up = 7. So "(missed 7 of 229 phases)", then the current formula would give RR = 94%

C) 3%x1127 = 33.81 RU = 34. So "(missed 34 of 1127 phases)", then the current formula would give RR = 94%
Same as for B, here. It works!

The only flaw I see is that "missed phases" are not the phases actually missed. But... who cares? We could hide that number and change C from "(missed 34 of 1127 phases)" into ("0 NMRs, 1 CDs in 1127 phases").

The important thing is that we could actually punish uncorrect/unfair people WHITHOUT screwing the RR of noobs or unlucky players who occasionally miss a turn: they can easily keep their RR above 90%.
Only seriously unreliable players would have a low RR. Only them would have issues joining new games.
I believe proportional penalty is the path.

Thoughts, anyone?

Nice idea.
We could change the name from missed to penality-phases (or hide them).
Any more ideas...? I would like to get some more feedback/ideas before I start coding.

I like Quaroz's proportional penalty for chronic CDers. In the long run, this will improve the quality of games by cutting down on what looks like a substantial CD rate here (almost every game I've played in has had at least one).

Sorry I wasn't taking on discussion last few days, as some people said (myself included) I got a bit too tense on the subject.
@Guaroz, totally agree with the work/idea, the numbers turn out to be quite well, especially considering that 6% on the first-turn CD (also for players with a lot of games) is huge. Everything above 80% is still very reliable imo. It seems that in this discussion the punishment could not be too hard in my opinion, whereas others are less strict about the reprimands. I doubt everybody will be convinced by either camps, though I want to endorse Guaroz' idea (if that's the procedure that's going to be followed afterwards). Apart from the fact that this proportionality in RR will take away the whole reason of denoting it in percentages, but that's not necessarily a problem, simply removing the %-sign and stating that 100 is maximum should be sufficient imo..
@Oli, are there as for now stats about this (say: how often do first-turn CDs for players with 100+ games occur and are there people repeating this practice often)? Especially because I think people abuse this loophole, but others defy that this happens. If researching this takes too long, then I'll give the others the benefit of the doubt.
The stats don't need to be shown imo, as long as this is mentioned as some kind of rule on the site.

No there are no stats about this.
The webdip code has (at the moment) no way to track CDs at all.

Oh, what a pity!
I was sure I did read some "left" in somebody's stats... maybe it was not this site.

The point is that I was about to propose 2 accessory tools. The reason is that many people don't play lots of games at once. For example, I usually play no more than 4-5 games contemporarily so a 63% RR would not be an issue for me. So:

1) Some play for good record. Adding a "left" in their stats (as they CD) would decrease their "win" or "draw" ratio.

2) Some plays for points. It would be important that when they CD they'd lose their bet. (Since I believe that now, as someone takes over the CD-Country, they have their points back. Right?)

These look 2 good accessory deterrent for CDs.

And yes, carpenter, you made some good point. User should reconsider worth of RR. A 92 would be a very good RR, while now it is not considered that good. Removing the %-sign would also help people remind that the scale has changed.

Further, my parameters were just an example. So, if you believe that 3% it too heavy for a CD, we could make it, say, 2.75% or 2.50%. No, problem...let's discuss it. Perhaps someone, who has just lost a game because his enemy has been advantaged by a CD, will propose 5% !! :-)

But I think it's only fair to ask as well: "are we going to apply this to every CD/NMR? Especially since we started this discussion only for the first-turn CD."
@Guaroz, I think I made my point of view clear: the higher the better, so I'll lead myself by more conservative/forgiving players on this site.
And I think 2 is already implemented, but points are not very important on this site.

The way I see it is a tradeoff between how fast the game moves on and how damaging NMRs are.
I understand that people want their games to move faster, but I believe that *not* having an NMR is much more important for enjoyable play. I would actually propose using the CD+extend for all phases, up to an extension of 24 hours for Movement phases, and 12 hours for retreat/builds. Sure, games will take longer when people NMR, but then the chance that a country will miss moves and change the game state is greatly reduced. The CD penalty could be increased too.

@Gobbledydook, then it is beneficial to join a game, communicate with people and then NMR the same turn, because 1. you dont like the kind of communication, or 2. Nobody is offering you a very good deal, or 3. One of said deals goes wrong and you don't like playing anymore with these players. Also purposely induced NMRs are also affected by this method (done it several times on other sites, I was planning to continue with that tactic to make other players err by not handing in my orders on time, everytime).

The Rating is more a proof of concept, not set in stone.
I wouldprefer to remove the % compltely and just make a text. "Very good", "Good", "Average", "Medicore", "Bad", "Very bad" Maybe add "Beginner" for people with phases under a certain threshold. Than you could set a minimum rating during game-creation.

Sounds good, but I believe that somewhere in the "help" section you should explain how it works. You may not know the exact value of your RR, just knowing it is "Very good", but you should know what happens to your capability of joining new games when you NMR or CD.

Another idea.
Each CD decrease your RR by 5-10%. After 100 phases played without a new CD all your CDs are cleared. If you have a new CD, the 100 phases to play reset and you have one more CD on your record.

I like this idea too. I would vote for a 10% reduction, as CD's are huge defaults and completely change the face of games. Once in a blue moon, one happens that is unavoidable, but for the most part, they are avoidable since you are missing more than one deadline.

I like this idea too. I'd say 3% for NMRs and 8% for CDs. But I'd say also at least 200-300 phases for reset. There are many people here that can do 100 phases in less than a month. Even faster if they play live games often.

BTW. Sometimes it happens that a player has NMR. Although he didn't save any order, the system doesn't even print the usual italic "Missed last phase" under their name. Probably because they did log in, but without saving anything. So I asked them about why they didn't save and sometime I got answers like this: "Uh, yea, sorry. I'm playing 15 games and yesterday I was rather busy. So I choose the games where I'm doing well, first. Sorry, I'm not very concentrate on this anymore."
Well, looks like those guys play 100 phases per week.

But my point is that I find rather strange that a user, even a retired 80-years-old man with no relatives no friends no job no books no TV & no other hobbies, is allowed to play that many games. How can they follow 15 games? They can't, so they choose their best games. They just log in for a few seconds the worst games, to avoid getting a "missed".
This happen more often in gunboats. Somebody even finalizes all "hold"!
So since somebody talked about quality games...

Start a discussion? I'd like to know if this looks strange only to me. Also, they could trick every phase-based RR system. Think for a minute. :)

Max limit on games?

I'm currently in 12 or 13 games and am focusing just fine on all of them, even the ones I am doing awful at. And I have classes and a social life to deal with as well. why limit those of us who can do such a thing just because of those who can't.
also, sometimes finalizing all holds is acceptable, rarely, but sometimes.

All players should know their limit. Surely butterhead, expert player, knows his own one.
Finalizing all holds is acceptable very rarely. Perhaps when you got only one unit left. But if you still got 5-6 units, it looks rather odd.

No, I wasn't thinking about a max limit.
Notice that games are not equally committing. Committing depends on several factors, mainly: phase length, press type, number of players and SCs in the map and... how talkative and careful you are when you're diploming!
I believe I could play 20 AncMed 4 days/phase Gunboats at once.
But, when I played a 24h/ph public-press-only Chaos, I could play ONLY IT! Too committing! Each time I logged in, there were tons of messages to read, and a lot to think on how/what to reply.
So there wouldn't be a "right" standard max limit. Each player must know his own one.

Actually, the issue I have is with those players who log in a game and log out 3 seconds later, just to avoid the "missed last turn". It's basically a hidden NMR. These players know their max limit, but they ignore it. With awareness. And they trick the RR penalty logging in for a few seconds, when they can't/don't have time/don't care to play.

Perhaps it is a coding issue, IDK if anything can be done about it.

haha yea my limit is(as I discovered) about the 20/21 mark... never doing that again...

At the moment there is a limit on how many games you can join.
For each 10% RR you can join one game. With a RR of 90%+ you can join as many games as you want.
Sure, every system can be tricked rather easy, but that doesn't mean we shouldn't try.

Shouldn't the RR be counting any missed phases whether you logged on during the phase or not?

Meaning if, as Guaroz said, "Actually, the issue I have is with those players who log in a game and log out 3 seconds later, just to avoid the "missed last turn."

They will get a RR penalty whether you get the missed last turn or you log on to avoid it.

@carpenter: There is a simple way to do it: if you leave a game in the first year (or whatever time seems appropriate), you may not join a new game for a certain time.

I've read all the comments above, and I feel like I do agree with some of the proposals about the penalities for chronical CD/NMR players
...but I think we're on the wrong path for the solution, or at least it's just a partial solution.

My opinion is that once a player loses 1phase, or worse, goes in CD the damage is already done and the game
is going to be strongly affected, REGARDLESS of how soon we're able to recover/substitute the player, and REGARDLESS how much we're going to punish the "guilty player".

Oli said there're no stats available about Civil Disorders, so I'm going to report my experience here as an example:
I've played 35 games on this site (not a huge number actually, but big enough to have an idea about what's going on).
Among 35, 18 are 1vs1, so I'll exclude them cause it's reasonable to assume that a 1vs1 game won't have CD issues;not sure,though!
I'm going to exclude 2 live games too, cause the dynamics/thoughts about live games are rather different.
We're talking about 15 games then.

5/15 (33%) games have had no CD.
5/15 (33%) games have had 1 CD.
3/15 (20%) games have had 2 CD.
2/15 (13%) games have had 3 CD.
10/15 (67%) games have had CD issues.
The average of (CD occurrence)/(Number of players per game) is 13% ,minimum zero, maximum 3/7 on a classic board...

14/15 games (93%) have had NMR issues.
(the only flawless game was a 1066 variant, three players...)

my diagnosis: CD/NMR is a plague and we should prevent it. Mods, forgive me if I'm too rough, but it's my opinion,and numbers argue in favour of it.

SO, the point is: why are we wasting our time in trying to fix a problem which is impossible to be fixed? Why can't we put our efforts in preventing the problem to happen?
I mean, like I meant a couple of months ago
http://www.vdiplomacy.com/forum.php?viewthread=18519#18519 :
why aren't we "reliable" player allowed to join a game with a reasonable assumption that the game itself won't be influenced by CD/NMR?
My request was simple: RR thresholds and number of phases played thresholds option to join a game.
Oli himself said this kind of change is obvious and easy to implement.

Summarizing, my proposal is: give a benefit to the reliable players, continueing punishing the "unreliable" players won't solve the problem:
the website has 2000 users, there will always be a newcomer who will "ruin" a running game.
Chronical CD/NMR players and multiaccounters don't really bother about RR.
It is very likely that newcomers too dont'bother about RR.

I do love this game, as well as this website, but the percentage of games affected is too much for my taste and now that I've brought my active games to the end I'm going to stay out of ongoing games until the problem is not solved.

Yes DL, Oli said he would have implemented thresholds options to join a game. But I guess that if he hasn't done it yet, it's because there's no use in it until we find a working and reliable RR system. That's what we're talking about.
With the current system, I could CD (double NMR) all my 7 current games and still have a 97%! So what's your point?

About newcomers, there is already a system (or at least I believe so: I didn't find it in the "help" section) that prevents noobs from joining more than 3 games until they have played a minimum number of phases (I don't know how many).

NMRs will always be. But don't forget that we can reduce them with an appropriate new RR system.
Both because it would be a real deterrence (and deterrents to work need to be known: have you ever seen Kubrick's "Dr.Strangelove"? LOL).
And because a good RR system would actually prevent chronical CD/NMR players from playing too many games.

"Optional restrictions to “unreliable” players" and "Thresholds" are discussions we can do later, if still needed. I'd have some ideas about it already.


__________________________________
http://www.youtube.com/watch?v=cmCKJi3CKGE
Dr.Strangelove: ".... all point of the Doomsday Machine... is lost... if you keep it a secret! Why didn't you tell the world?!?"

I'll add the option to set a min rating for your games in the next few weeks (had a really busy RL, and many new variants needed my attention too.
But as Guaroz alreadymentioned. The CDs are a different matter than phases missed. I'd like to monitor both in a single rating to keep it simple, but surely a CD is much worse than a forgoten phase.

Sounds good Oli -- thanks!