what should we do than?

Remember that after this round groups all games will have to end at 1 month period.

And, cant people already share their password or guide play or have fake accounts despite this rule?

with warning system, can't teams make deals with each other to not give each other warnings

I've seen this happen already

How, praytell, have you seen this when you aren't involved in the tournament?

http://WarLight.net/MultiPlayer.aspx?GameID=2861812

Perhaps a better system would be to use the "Speed" feature and check everyone's average turn length (it shows last 10 turns) and average it to come up with a multiplier?

---

thats not a bad idea
the teams should be accounted for the slower player than?

---

How about the following:

• Copy each players "Playing speed" list
  
• For each number subtract 24 hours (if it becomes negative, consider it zero; we only want to give a mild penalty to players going over 24 hours, not give a bonus for playing really fast, right?)
  
• Do something sensible to create a "fair" average
  
• Use that average (maybe after some scaling) in the same way you wanted to use the 3% penalties.
  

Let me know if you need help on the formulas. Just tell me how it should work (current assumption: spending 1 minute or 24 hours is "the same"; do you want to mostly penalize going a little bit over 24 hours all the time, or do you want to mostly penalize going far over the 24 hours a single time?) and I'll whip something up.


I like this system a lot more:

• It's less work (less bookkeeping)
  
• If teams decide to make an effort to finish a game, they will benefit (since only the last ten turns count)
  
• I'd feel guilty about issuing a warning like originally proposed; this system puts everyone in control of their own "scoring modifier"
  

Yes, I'm aware really nasty opponents can still game the system a bit (I won't go into details, for the same reason as Yeon), but it only applies to Round 1 (and in later rounds when a team sets themselves up for it) so I think it's not too much of an issue.

with warning system, can't teams make deals with each other to not give each other warnings

Is that a problem then? If each team plays "fast enough" there won't be warnings either. When we reach the deadline, the games can still be ended "by formula", there just won't be any "+3%" penalties that have to be taken into account. If teams agree on this, why not let them?

in my view this is a bad idea

orinocoflows, there's a lot of ideas in this thread; which one do you consider a bad idea? Secondly, could you explain why you think it's a bad idea? Just saying it's bad is not very helpful and not very convincing either... Lastly, do you have a better suggestion?

I second RvW's idea - the original idea of issuing repeated "warnings" seemed unnecessarily antagonistic and I think the spirit of friendly competition would have been ruined by it. RvW, please go ahead and propose a concrete formula so we have something to decide on.

RvW: For your idea, perhaps we don't need an average at all? Perhaps just take the slowest player's record and use that.

I do think that system is more gameable than the initial one, but it's much nicer to
1) Just focus on the last ten turns
and
2) Avoid the antagonistic warnings

So I think it's better.

I'd like to propose a variation on RvW's idea, though. If we're willing to do a little work, then perhaps use a formula like this:


For each of the 10 last turns:
Take the number of hours the slowest player (P) used that turn = X
Take the number of hours the slowest player who isn't on Ps team used = Y

If X is less than 24 then X becomes 24
If Y is less than 24 then Y becomes 24

Z=X-Y

Then Z is the number of hours beyond 24h the slowest team delayed the game. Add this for each team, and allow a mimimal grace (say, 12 or 24 hours total for the last 10 turns). Hours beyond this are added with multiplier to the opponent's score (0.5% per hour?).

Pros:
* It's less gameable (as far as I can see): Delay caused = points lost
* We avoid the warnings
* There is no work to be done underway

Cons:
* It does mean a bit of administrative work when game ends.

Thoughts?

Yeon and RvW,
Would it be somewhat simpler if we did the following:

1. For each of the last ten turns, find the slowest player on each turn.
2. On each turn, the slowest player's team is penalized for every hour over 24 hours that player took to move. No penalty is applied that turn if everyone on both teams moved within 24 hours.
3. Add up the penalties for both teams over the ten turns and take the difference to get the total penalty for the slower team. Let that number be P.
4. If P > 48 hours (the grace period, equal to a team using the entire three days once during the last ten turns), then that team's score in the adjudication is (1 - (P-48)/432)*(income*2 + number of total armies). For the faster team it's simply the unadjusted formula: (income*2 + number of total armies). If P < 48 hours then no penalty is applied to either team.

For example, let's say a game was created on Aug. 1, everyone joined by Aug. 2, and slower team used the full three days for every turn, so that by the end of Aug. 31 exactly ten turns had been played. Then P = 10*48 = 480 so the slower team's score would be divided by half: 1 - (480-48)/864 = 0.5. Is that reasonable?

Sorry, screwed up the formula the first time! (It's correct in the example.) #4 should read:

4. If P > 48 hours (the grace period, equal to a team using the entire three days once during the last ten turns), then that team's score in the adjudication is (1 - (P-48)/864)*(income*2 + number of total armies). For the faster team it's simply the unadjusted formula: (income*2 + number of total armies). If P < 48 hours then no penalty is applied to either team.

I am pretty happy that the several guys now are helping me considerably,
I am likeing this discussion, it takes out the weight in my shoulders a little bit.

I am 100% ok with the formula and criteria you guys end up realizing is the best way to do it.

Aziridine: I think most of what you said is reasonable, but I don't think it fair if slowest player on team A uses 37 hours and slowest player on team B uses 38 hours, and then team B is punished for delaying the game by 14 hours. Don't you think the fair emount would be one hour?

i'm not part of the nations cup, but i think ( the first post by dunga) is a bad idead

warlight is just a game, and people are meant to have fun, this will be forcing people to play at least once a day, which definitely isn't fair because people like to socialize and do other things on holidays, rather then just sit at home and play wl

But let's say team A committed orders 24 hours ago but team B still hasn't moved yet, and now team A has thought things through a little more and wants to change their orders. If they do so I don't think the penalty for the team B should go from 24 hours to 0. What do others think?
Nike, would the 48-hour grace period over the last 10 turns be enough to satisfy your concern? Or how much longer do you think it should be?

I definitely think the advantage should be reset to zero if they send revised orders, since new orders are indiscernible from revised orders. And I think it'd be very ugly if one team can bank lots of hours over theoretically being one second faster.

As I see it we're trying to punish teams that drag games out on purpose. What the other team does while it waits for the slower team shouldn't change how the slower team is punished. Also I think this is slightly easier to calculate than your alternative. But all in all it's a minor point and I'll concede it for the sake of having a rule in place by Monday. So replace my #1 and #2 with your original formula.

Is the rest of my formula acceptable, including the 48-hour grace period? I think the original example I gave was pretty extreme (one team taking 20 days longer than the other team over a 30-day period) which led to the drastic point reduction. If a team uses only 48 hours beyond the grace period (so they used 96 hours or 4 days more than the other team) then the multiplier becomes 1 - (96-48)/864 = 0.944, which seems fair enough to me.

I was also looking at taking the maximum, but unfortunately there's a (solvable, but still) practical problem with it:
WL only lists turns actually taken, so if player A and player B have been playing at "the same speed", until player B suddenly stopped committing orders, the list will look like this:

Player A    Player B
65          36
23          65
24          23
64          24
13          64
 2          13
 4           2
12           4
18          12

So when we process these stats, we'd have to be very careful about which turns match up. And, how do we factor in the time we've been waiting for player B? In the example it could be anything from 18 hours (since it must be longer than player A; the turn starts at the same time for everyone) and 72 hours (since autoboot would've kicked player B at that moment).

Proposal:

• Add the "we have been waiting for x hours" timer as an extra data point to the playing speeds of all players who have not committed orders for the "current" turn when the round ends (both for fairness and to make it more difficult to forget "lining stuff up").
  
• Only look at the most recent eight turns. This is to prevent things from getting messed up (data "going missing") if the last player commits orders between the round ending and someone (Dunga?) having time to look it up and write it down.
  
• Of course, the "we have been waiting for x hours" value has to be corrected as well for the amount of time between the round ending and values being checked.
  
• This also buys some extra time in case teams disagree; more time before data disappears which is needed to verify the rules were applied correctly.
  
• It looks like you already have an idea how to process this "normalized" data further; terribly sorry, don't have time today to look into the details. :(

---

warlight is just a game, and people are meant to have fun, this will be forcing people to play at least once a day, which definitely isn't fair because people like to socialize and do other things on holidays, rather then just sit at home and play wl

In all fairness, this rule only applies to the Nations Cup; people knew what they were getting into before signing up...

Excellent points RvW - I like all your ideas.
But honestly, the more complicated this gets the more I dread actually doing the adjudications. Are we going to apply this to the elimination rounds as well? If vacations aren't going to be honoured at that point I think it would be better to just play the games out - there will be fewer results to wait on anyway.