Players can't join games already because of passwords. I don't think RR will become that popular, tbh, that someone actually can't join games. If it does, the problem can be addressed then. Bridges and all that.

Yeah. Typing out my last post, I agreed more and more with you. A major point is that RR is just points for logging in to a board game website on time. It doesn't really matter in the big picture of life.

"If RR is at the discretion of the mods then it is not particularly useful metric, imho."

Absolutely. Ideally, the only way to improve your RR is by being reliable in games.

Well, I for one wouldn't want to see it implemented until there's a way to differentiate between people who CDed a couple of years ago and those who did so yesterday, but if that's done, I think it would be easier to make a game with, say, 101 and high RR requirements than to find people on the forum if you want reliable, experienced players.

You're right though, it won't be that popular. Just quite useful, *if* the RR has a bit more sense to it.

Without knowing the algorithm, it's hard to make specific suggestions, but a way for players to atone for CDs by, say, taking over other CDs could be a way for a couple CDs not to ruin a player while also encouraging service to the site.

"Well, I for one wouldn't want to see it implemented until there's a way to differentiate between people who CDed a couple of years ago and those who did so yesterday,"

Agreed. There should be a weighting function in which old CDs are less impactful than new ones.

Of course, it's a lot easier for me to come up with ideas than for ATC to implement them.

I'd like fixed RR requirements for live games. It'd be another step to improving the live scene.

Full-length games seem like a better way for A) People to fix their poor RR's and B) where people can explain why they have poor RR's if they have legitimate reasons, and thus be arranged and accepted accordingly.

Abge: I'm against CD takeovers affecting RR for the same reason that intervention with the numbers is bad. Taking over a CD doesn't mean you're any more reliable, it just means that you joined a game.

I agree that it's worth incentivising service to the site, and I think there are other things that we can do to incentivise CD takeovers. We halved the price of all CD takeovers a few months ago, and have seen a large drop in open positions since then. Possibly protecting players from GR hits would help too, but we'd need to think carefully about how to do that without compromising GR.

It's after 2am here, so I'm going to go to bed before I end up arguing with myself again :)

Why not implement a tool similar to what fantasy basketball uses? A Reliability Rating based on the last 90 days, last 180 days, last year, and last three years? This would allow folks a better idea of whether a poor rating was tied to a particular point in someone's life or whether their rating has been steady across time?

Works for me but that's a bit tricky when setting up requirements. You'd have 4 seperate requirements.

Unless you take the average or something, which is a different way to make recent CDs weigh heavier than old ones. It's certainly something to consider when trying to figure out how best to implement weighted CDs changing over time, I'd say..

Yes to gated RR games. I would just love it, (though, tbh my RR isn't that high, had some "accidents" on the way). I also love the idea ATC had about RR being comprised of last X games, or games from last X years.

I myself has never intentionally abandoned a game (except for maybe like my first game here when I didn't know any better). There is no excuse to CD. The game is already ruined? Well, don't ruin it even more for the other players. You have like 1 or 2 SCs and will SURELY die in the couple of next turns? Well, first of all there is nothing sure about the future, and even a 1-2SC nation can have a great impact, that you deprive the game of by CDing. As I mentioned, I have never decided to just leave a game, there was always a reason - blackouts, missing a bus and getting home late for the turn, losing my phone on holidays, and still, if someone looks at my RR and says "well, 90% just won't make it, I don't want to take those chances he'll CD - I don't want him in my game" he has every right to do that, and I hope soon he will have a way to do it as well. Because whatever reasons you try to give to justify you CDing, it just means that for the same reasons you'll CD in the next game, and kids - CDing is bad no matter what.
@ATC, you should take a little break from coding and shoot a public service announcement with puppets and stuff about why CDing is wrong.

I like the idea of a RR, but I agree with steephie that the system needs to be adjusted. My first concern is that it treats all CDs the same, whether they happen in 1901 or 1925. Obviously the CD has more of a negative impact on the game the earlier it is and the more powerful the player is in the game. If the goal of the system is to improve reliability when it matters, CDs and NMRs should be penalized most at the start of games (especially with regard to neutral centers) and least at the end of games. If Austria, Russia, and Italy are holding a stalemate line against a powerful England, can we blame Turkey for sitting idle? Also CDs and NMRs when a player only has one center should be given less weight, since what can the player do in that situation anyway?

As you can see from my RR (87%), I'm fairly reliable. Just curious, what do you guys think would be popular standards if RR became commonplace? 60%? 75%? 80%? 90%? 95%?

Also I can sympathize with steephie. RR seems to be hurting him ex post facto, when he had no idea that he would be penalized for quitting a game that in his opinion wasn't worth playing since something else had already ruined it.

There are simply cases in which the entire board doesn't give a shit. No one is harmed by CDing then and you are harmed by playing on.

That is, unless you decide to add a rating later and attach game requirements to it.

That message was directed at yassem for the record, thanks for the support Brankl.

What if the gate could be customized, not only for actual RR, but for RR over a certain period, say last 20/50/100 games or last 6/12/24 months? Or any arbitrary number?

I know that means there'd be live calculations based on historical data instead of an RR field populated by a periodic calculation, but would that create a huge server load?

Then the steephie's of the world could adjust their criteria accordingly and set up games which work for them.

How exactly does playing on harms you? ._. It leaves emotional scarring or something?

I think Jeff's idea is a better solution than a one-size-fits all, but I also think that there should be a limited set of standardized options like Eadan suggested so that unreliable players have less ability to skew the data and confuse innocent victims. (Unreliable players shouldn't be able to make games that have a reliability gate for the last 123 days when they CDed 5 games 124 days ago). I also disagree with others who suggest that mods should stay out of the equation. Yes, there needs to be independent verification of reliability, but in highly unusual circumstances players need to feel that someone cares about their situation and is willing to help them.

Harmed is a bit of an overstatement, but that time could be spent earning money, doing homework, working on a project or having fun. Any time you spend on a ruined game no one cares about is wasted time. Any energy you invest in it, wasted energy.

Yassem, I hope this is a site that cares about its memebers' time. Wasting time on a game that should be cancelled just to keep a high RR is still wasting time. Hopefully this new feature doesn't encourage players to waste their time instead of doing something productive or meaningful.

I also think that a single RR (or at least, a singe RR used for RR gating) is enough, if it will include the element of not involving all the games, but rather X last. It's reliable. It should be quite easy to implement for our God ATC. It will be easy to set up the game with RR. It will be awesome.

Stephanie, you already committed to the game when you joined. The fact that the game sucks doesn't change the fact that prior to finding that out you thought you had enough time to play it. The fact that you make justifications like this is exactly why your RR sucks, and why for example I wouldn't want to play with you (apart from that king on the hill thing, it looks interesting).

On mod intervention, the problem is that any good Diplomacy player can probably convince the mods that something horrible happened to him, should he want to. I don't care too much either way though. The mods are smart enough to keep things under control.

Another option is to put a higher weight on high stakes games, or if possible, on games with other reliable players. For games in the database, I think high stakes is a rough approximation of other players' reliability. Higher stakes games just don't have as many CDs and NMRs, and thus are unlikely to be those "games not worth playing" (that should have been cancelled) that steephie mentioned.

I see the benefit of no mod intervention, and I'm not entirely against it, but I do think that special circumstances need to be dealt with. And by mod intervention I mean not counting certain games (that should have been cancelled) into the calculation of RR. Not necessarily claiming that something terrible happened, though I do believe a little sympathy would go a long way. In general though, the system needs to be independent of the mods. I'm only talking about very rare circumstances (or absolutely terrible games with multiple CDs), and in the end it doesn't make a big difference to me.

@yassem: I had enough time to play it, so I can, for instance play another game in that time, one people care about.

I don't do this anymore (I only play games where no one CDs these days anyway), but if everyone on the board is cool with me leaving the ruined game, so am I.

Now there's RR of course, but that seems like the sole reason to continue in such cases.

Don't act like this is general practice for me. The last time was over two years ago. The problem is that you may very well be less reliable than me, yet my rating is shit. It simply isn't accurately showing my reliability.

Steephie's rating shouldn't be automatically adjusted, but perhaps a mod could look at specific examples steephie believes are unfairly hurting his rating and make a decision after having considered what exactly happened.

Which one of ATC's example (RR from last two-years) would solve, In the other case, just play 50 games without CDing : )

That works well as long as steephie can join 50 games with his low RR. Not a problem now, but it may be in the future.

Well, he isn't the only one with low RR.
Another idea though, RR could be improved a little extra if you replace someone. Could that maybe work?

I think it was ATC who mentioned earlier that replacing someone does not mean that you are more reliable, just that you want to play another game.