Assume you want captures only to decide the game - how should the personal scoring look like?

I am going to set up polls to answer simple questions to decide parameters of the scoring.

The questions will be:

Which success rate must a capper have to be of value for his team, that is, to get a positive personal score?
(0% to 50%)

How many flag carrier kills with returns are as good as a successful capture, in terms of contribution to the team?
(3, 4, 5, 7, 10, 20)

How much credit to give the flag carrier killer compared to the returner?
(reward the returner only, 1/5, 1/3, 1/2, 2/3, 4/5, reward the FC killer only)

These polls with their results should give a general idea on how to tune the personal scoring system.

ONLY VOTE if you think of the implications of your choice.

regular_kills*A + fckills*B + capture_value - losing_flag_value + picking_up_flag_value + returning_flag_value - teamkills*C - suicides*D

The losing flag punishment should depend on the other values and constants.

bonus_if_returned_by_fckiller value can be added too, but not necessary.

Don't you think this is the very POINT of this thread? Namely, getting some input in clear terms, from which to calculate the points for the different actions?

The punishment should be adequate in comparison to all other scores, so you can't decide the punishment's value basing it only on the capture score, what you are now trying to do. What I mean, punishment value should depend on return_value + fckill_value*B too. If you make those two higher/lower, punishment should be decreased/increased accordingly.

IMO, the best way would be if we could have statististics for the current average caps/pickups ratio and base the punishment on that and not on what pops out from the head.

Too bad i can no longer do any statistics, as my server lost its popularity.

Nexuiz players instead want 20 players hook CTF on 10 square meters maps.

You have (or did you lose?) older statistics. I doubt that average changed a lot.

I don't keep log files forever, neither am I allowed to by the laws.

Maybe we could use dcc stats then.

It is simply illegal to even have such old logs, even more, to use them.

Code: Select all

0% (no negative scores for any capture attempts)            
    15%    [ 2 ]
10%            
    15%    [ 2 ]
20%            
    30%    [ 4 ]
30%            
    23%    [ 3 ]
40%            
    7%    [ 1 ]
50% (so every second cap attempt must be successful, or there is negative score)            
    7%    [ 1 ]


Will reset the poll, and make the next one. 20% is the clear winner, maybe 25% will be viable too.

Logs of ip's (in your country) and not the stats db.

The server log still contains the nicknames, which can be argued to be personal data.

Nonsense, because XSAX can be me or YOU anytime he wishes. Or I can be him too

Even IP is not proof (dynamic ips, infected proxified pcs, public access terminals, ip spoofing)

I really am not sure i understand this poll...i think i do for the most part though.

but for example a (flag kill) should take 7x more to = the same amount of points for a flag-capture is what i'd propose...

so 7x FC kills = 1 flag capture.

im not sure about the WITH flag return, i was thinking of just if you kill the FC since you might not always be the one to return it or be able to.

Just whilst im here id propose a system like this (for personal scores).

Flag capture = 21 points
Flag kill = 3 points
Flag return = 3 points
Frag = 1 point

or something very very similar, thats just the basic stuff i can think of..no idea if you are going to include assists...

Imagine you get 100 points for capture.The question is how many points you should give for return?

Alien wrote:Imagine you get 100 points for capture.The question is how many points you should give for return?

Ok I edited my post, i explain it there.

Wheres the option for none?

There is no option NONE, as this is for personal scoring, NOT for team scores. Team scores being a capture counter is unharmed by this.

And intentionally this only covers the case when the capturer also returns. A second poll will decide how to distribute the points between capturer and returner.

From this poll, we get:

g_ctf_flagscore_pickup_base + g_ctf_flagscore_capture = 25,
g_ctf_flagscore_pickup_base - g_ctf_flagpenalty_drop - g_ctf_flagpenalty_returned = -6,
g_ctf_flagscore_kill + g_ctf_flagscore_return = 8

Code: Select all

(%i26) solve([
  g_ctf_flagscore_pickup_base + g_ctf_flagscore_capture = 25, # 25 points per cap, generally agreed upon
  g_ctf_flagscore_pickup_base - g_ctf_flagpenalty_drop - g_ctf_flagpenalty_returned = -6, # fits the success ratio of "slightly more than 20%" from this poll
  g_ctf_flagscore_kill + g_ctf_flagscore_return = 8, # fits the defense points from this poll
  g_ctf_flagscore_kill = 3, # arbitrary by me
  g_ctf_flagscore_pickup_dropped = 1, # arbitrary by me
  g_ctf_flagscore_pickup_base=0, # arbitrary by me
  g_ctf_flagpenalty_returned=0 # no return penalty (we'll do drop penalty instead)
]);
(%o26) [[
  g_ctf_flagscore_pickup_dropped = 1,
  g_ctf_flagscore_return = 5,
  g_ctf_flagscore_kill = 3,
  g_ctf_flagpenalty_returned = 0,
  g_ctf_flagpenalty_drop = 6,
  g_ctf_flagscore_pickup_base = 0,
  g_ctf_flagscore_capture = 25
]]


This system gives 6 points drop penalty, and it can't really be avoided. However, removing 6 points at once is harsh. So let's make it 3 points when taking the flag ([-z-]'s deposit), 1 when dropping the flag, and 2 when it gets returned. Then retaking the flag after one died zeroes out the drop penalty, and you get another chance as if nothing happened.

Let's add [-z-]'s deposit...

Code: Select all

  g_ctf_flagscore_pickup_dropped = 1,
  g_ctf_flagscore_return = 5,
  g_ctf_flagscore_kill = 3,
  g_ctf_flagpenalty_returned = 2,
  g_ctf_flagpenalty_drop = 1,
  g_ctf_flagscore_pickup_base = -3,
  g_ctf_flagscore_capture = 28


Now the number 28 looks a bit odd. It basically is 25 points, and undoing the 3 points deposit. We got our scoring system.

Of course, we can freely shift points between return and kill, but I'll start with these values, as I believe the returner does the more risky work (as he does have to enter enemy lines to retrieve the flag).

It looks good but I think the drop should be at least 2 points... again for pyschological reasons. 1 point just doesn't seem too bad from a personal standpoint. Teamwise, it hurts (with the kill + return) but I doubt many players will make that connection.

Well, the loss of 2 for a return is a higher incentive to get the flag back (to avoid that loss). Alternatively, one could use 2 drop penalty, 1 return penalty, and 2 score for picking up again.

But the return is the time when the flag is actually lost and the round "decided", after all.