Slitherine

Hi guys

if you can post your results here we can check the system is scoring them correctly. What we want to know is if the game had a clear winner - i.e. one side was completely routed and the game ended or if it was a draw where both sides had troops remaining. If there was a clear winner the % losses of the winner. If it was a draw the % losses of both side.

I can see one game has completed already and 3 are under way!

OK.

Let's see:

1st Round Clastium 222BC

tofman04 (Roman) vs maximvs (Gallic) 9/9 vs 6/12

Defeat (Losses: 100% vs 50%)

Iain,

Just looked at the results table thus far and:


Now I'm guessing here, but I'd assume the 0-0 game means it hasn't started yet;
The 15-15 games means the game is running, but equal in break-points;
And the 0-30 game means it's started and epos would gain 30pts if the game was to end right now.

So does this mean that you can 'watch' the games in real-time during the tournament?
If so.....Very cool

Lars

Yes you can see games in progress but we were only doing this for the beta. We can look at ways to keep that option available as long as people can see which games are complete and which are not.

1st Round
EricS (Gallic) vs NFTaylor (Roman); in-game victory display shows 0/9 vs 0/12
NFTaylor ran out of time

1st Round.

CheerfullyInsane (Roman) vs epos (Gaul); Time-out victory for me.
Losses stated as 44% vs. 17%

OK round 2 is up!

It looks like jimcdaw & stonebraveheart have been cut from the draw. This means these players did not play a single turn in their matches & ran out of time. If this is not true for either of you let me know!

iainmcneil, petirogue and iandavidsmith all got max wins which suggests their games timed out withotu them taking any losses and this all appears correct as I had no losses and the other 2 were against people who the system thinks never played a move.

Here is what it should do - I havent had time to check if its calcualting as intended

If player1 has run out of time ->
Score1 = (ScorePlayer1 / 2)% * MAX_POINTS;
Score2 = MAX_POINTS - Score1;
If player2 has run out of time ->
Score2 = ((100-ScorePlayer1) / 2)% * MAX_POINTS;
Score1 = MAX_POINTS - Score2;

Here's how the game calculates the scores to send to the server, both during the game and at the end:

Let Player1Losses be a real number between 0 and 1.
Let Player2Losses be a real number between 0 and 1.

* If Player 1 resigned:
ScorePlayer1 = 0

* If Player 2 resigned:
ScorePlayer1 = 1

* If the game is over, and Player1Losses = 1, and Player2Losses < 1:
ScorePlayer1 = Player2Losses / 2

* If the game is over, and Player2Losses = 1, and Player1Losses < 1:
ScorePlayer1 = 1 – (Player1Losses / 2)

* If Player1Losses = 0 and Player2Losses = 0:
ScorePlayer1 = 0.5

* Otherwise:
ScorePlayer1 = Player2Losses / (Player1Losses + Player2Losses)

Iain

I like this scoring system much better than the first. One question though :-

Do I understand this correctly
In our game my loss was 2 yours was zero when the game timed out.
Had the game timed out without that routed unit both scores would be zero.

Had that happened then the condition that applies is
* If Player1Losses = 0 and Player2Losses = 0:
ScorePlayer1 = 0.5.?

If that is the case both receive 1/2 max points.
So there is no penalty for time out.

I think that might be a point in the scoring system with an unforseen consequence as in real competition being timed out could be gamed to benefit by someone loosing heavily or finding themselves manouvered into a significant disadvantage. They would thereby fix their point loss and so score at current value.
I am sure no one would do that but to remove temptation;
Have you considered already a penalty modifier for a player timed out whether it be calculated or agreed for any given tournament e.g. In this one -5 points for being timed out in a game?

Of course that would make my current score -5

No if we both had 0 and it timed out I would still have 30 0. Timing out effectively means you lose the game if it is not completed by the end of the time limit. E.g If I had lost 50% of my army and you had lost nothing, if you ran out of time it would be as if you had lost 100% and I had lost 50% which I think gives a score of about 22 8. This is better than just giving 30 0 to the one who did not time out as it means incomplete games still give points to the player who timed out but just less if that makes sense. Its less harsh than it used to be.

So this condition
* If Player1Losses = 0 and Player2Losses = 0:
ScorePlayer1 = 0.5
Only applies to games in progress - not irrespective of completion?

That makes sense now.

2nd Round Dertosa 215BC

maximvs (Roman victory) vs Cheerfully Insane (Carthage defeat) 3/14 vs 17/15

Note: The table shows a timeout for maximvs though, is this an error? I stayed up until about 23:30 last night waiting to see if I would get another turn. Then went to bed and then logged on again at approx 08:30 to find it was my turn but it said 0 time left. Went on to score a nice win though. Good game 'CI'

Time runs down while it is your turn so this could well have timed out over night. We have avery short time limit in this test so most games are running out of time but thats one of the things we wanted to test.

Have I got this right? If my opponent sends me a move at 11pm, I automatically lose unless I can make my own move before 7am?

Niall

Worse still, if your competitor sends his move at 11:59, you have 1 minute to open the game and finish your turn!?!

Mind you, this is what the testing is all about. It looks like I should have stayed up waiting for my competitor's turn to appear, rather than going to bed. I think this situation is only being highlighted because we are running a quick turnaround. I don't imagine that many tournaments that run after the beta testing will run to such a small window.

Probably not.
If memory serves, the last test of the system had each game run over nearly a week, with 3 days allocated to each player.
Don't think we had any time-outs in that one, except for those that dropped out altogether.

Lars

2nd Round
petirouge (Roman) vs EricS (Carthaginian); in-game victory display shows 11/14 vs 3/15
petirouge ran out of time

Triarius wrote:So this condition
* If Player1Losses = 0 and Player2Losses = 0:
ScorePlayer1 = 0.5
Only applies to games in progress - not irrespective of completion?

Yes, that condition will be be modified to penalise the side that ran out of time at the end of the game. (Except when the game is played to the end with neither side side resigning, and neither side having any losses, but that will be rare!)

2nd Round result

epos (Roman) vs tofman04 (Carthaginian victory) ; in-game victory display shows 8/14 vs 4/15

epos ran out of time (Losses 57% vs 27%)

3rd Round Corinth 146BC

1st game: maximvs (Achaean victory) vs triarius (Roman defeat) 14/19 vs 25/24

2nd game: maximvs (Roman victory) vs triarius (Achaean defeat) 23/24 vs 23/19 - that was SO close at the end!

Thanks for helping us test this.

Congrats to toffman for winning this beta tournament!

The next tournament will be on the live server!