Slitherine

I prefer a simpler system. The one used at Warfare worked well and only needs to be mathematically adjusted to make a 25:0 checksum.

I think it already has enough advantage for defeating the enemy. Giving a higher proportion of the checksum for breaking the enemy army also inevitably gives a higher proportion of the checksum for not being broken. This can lead to deliberate slow play and can cause a great deal of displeasure.

I prefer the system used at Warfare. This effectively gives +4/-4 for getting that last AP. (in 32:0).

Even if you decided you want a slightly larger bonus, this can still be done in the one table by adjusting the points retained for being 1 AP off defeat.

rbodleyscott wrote:I prefer a simpler system. The one used at Warfare worked well and only needs to be mathematically adjusted to make a 25:0 checksum.

In order to help those of us poor souls who could not make it to Warfare, what system was used there?

Marc

Your army is worth 10 points.
You give 1 point to your opponent for each full 1/10 of your AP you loose - so that to loose the full 10 points you have to break.
As this is dependant only on the number of BGs in your army this can be calculated in advance and listed with out order of march, so no calculators needed.

Bonus points (winner/looser)
5/0 - Broke opponent's army without own army breaking.
3/3 - Both armies broke (works out as 13:13)
4/1 - Neither army broke, winner has gained at least 7 points from opponent, and gained at least 3 more than they lost.
3/2 - Neither army broke, winner has gained at least 5 points from opponent, and gained at least 2 more than they lost.
3/2 - Neither army broke, winner has gained at least 8 points from opponent, and gained one more than they lost.
2/2 - Any other result.


Examples
Closest outright victory - 16:9 {10+1+5 : 0+9+0}
Closest 25 point draw - 15:10 {5+7+3 : 5+3+2}
Most extreme 25 point draw - 23:2 {9+10+4 : 1+0+1}
Most extreme 24 point draw - 16:8 {4+10+2 : 6+0+2}

Note: A one point difference may be a very minor difference in losses - as 10/17 vs 10/16 is a one point difference but the actual losses are a close as they can be, the difference in losses is only showing because they are on opposite sides of a rounding boundary.
There is an argument for doing it as 1 point per 1/11th and discarding the first point of difference.
i.e. calculate the absolute difference in points gained, reduce this by one minimum 0, and the winner gets 10 + x, and the looses 10 - x.

Also while it isn't zero sum the only way for both players to gain points by adjusting the final score is to convert a 12:12 incomplete game to a 13:13 mutual break.

sagji wrote:Your army is worth 10 points.
You give 1 point to your opponent for each full 1/10 of your AP you loose - so that to loose the full 10 points you have to break.
As this is dependant only on the number of BGs in your army this can be calculated in advance and listed with out order of march, so no calculators needed.

<snip>

My initial concern is that if I make an army with >10 BGs I can lose a BG with no consequence. Would this scoring system skew army construction toward larger armies in order to reduce the consequences of lost BGs?

Marc

babyshark wrote:

rbodleyscott wrote:I prefer a simpler system. The one used at Warfare worked well and only needs to be mathematically adjusted to make a 25:0 checksum.

In order to help those of us poor souls who could not make it to Warfare, what system was used there?

Marc

I would post it, but I don't have a copy. (It is not the one Alan [Sagji] describes).

babyshark wrote:

sagji wrote:Your army is worth 10 points.
You give 1 point to your opponent for each full 1/10 of your AP you loose - so that to loose the full 10 points you have to break.
As this is dependant only on the number of BGs in your army this can be calculated in advance and listed with out order of march, so no calculators needed.

<snip>

My initial concern is that if I make an army with >10 BGs I can lose a BG with no consequence. Would this scoring system skew army construction toward larger armies in order to reduce the consequences of lost BGs?

Marc

If you have 15 BGs then there are 5 APs you can loose that don't alter the score. The only way to improve this is to increase the points for the army, but there will allways be some magic points except when the number of BGs is a factor of the number of points - if we use 100 points then for 15 BGs there are 5 AP that cost 6, and 10 that cost 7.
It would be possible to do, but it takes more space for the tables to work it out - the 10 point system requires a 9x9 table for any army size, the 100 point system needs 124 values to cover armies of size 11-19.
Instead of using 100 points I would do it as 10 points to 1 decimal place.

An alternative is for each game to be 100 points - making each army 40, and the bonus out of 20.

FWIW the system used at Warfare (and Brtitcon which was similar) worked by using a table to give each player a value for VP kept and VP lost.

Essentially I think it worked on a 12 point scale (but may be wrong) where if you had 12 BG then each BG you lost would cost you a VP and give your opponent a VP. The last AP you lost cost more as it broke your army and there was an army break bonus such that with a broken army you kept 0 points and lost 16.

If you had more than 12 BG's then the VP lost were distributed proportionaly (I am not sure of the rounding) but there was definitely no possibility of losing 5 BG's and not losing any VP. I suspect that you probably had to have 18 BG or more to be able to lose one and not take a hit.

I like to think of it as almost a Duckworth Lewis style scoring system with a win bonus.

This is the Britcon scoresheet but the Warfare one was on a finer grain so there was a column for each army size rather than pairing them.

hammy wrote:This is the Britcon scoresheet but the Warfare one was on a finer grain so there was a column for each army size rather than pairing them.

Which is an important improvement. Otherwise there is a significant (inappropriate) advantage to having an army of the lower of each pair of total BGs.

(And I deliberately took an army with 11 BGs to Warfare in the hopes that the same system would be used again. It wasn't. Curses, foiled again).

hammy wrote:
I like to think of it as almost a Duckworth Lewis style scoring system with a win bonus.

Guaranteed to endear it to anyone looking for a simple system then

I think most of my results in Warfare were fair to both players. One thing I didn't like was losing points because my army was broken. Quite often I could have remained unbroken if I had played for a losing draw (by withdrawing troops from combat), but in those circumstances I went for the extra enemy BGs and that cost me in overall tournament points. This doesn't seem right somehow.

I'd prefer a system that was checksum at its core and then rewarded players for breaking opponents over that total, but with no penalty for those who had their armies broken. They are already going to be penalised for losing the game, so why penalise them more for being agressive? This type of scoring also encourages players to go for results and finish games earlier (=more booze and time to see other games and market stalls).

I'd prefer a system that was checksum at its core and then rewarded players for breaking opponents over that total, but with no penalty for those who had their armies broken.

The problem is that any checksum system must penalise players who lose - if it rewards players who win.
score for playr A + score for player B = 32 (or 20 or wnatever)

I suppose the question is - Do we need to use a checkpoint sytem.
The 3-2-1-0 system isn't a checkpoint system and has plenty of support.

We could easily share 20 points between the players and give a 5pt bonus for breaking your opponent.
A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

terrys wrote:
The problem is that any checksum system must penalise players who lose - if it rewards players who win.
score for playr A + score for player B = 32 (or 20 or wnatever)

I suppose the question is - Do we need to use a checkpoint sytem.

Well we know JD is paranoid so wants this as a basic requirement but it'd be worth asking a wide selection of players whether they feel it is necessary as ultimately they are the users. Worth reminding them that it would be hoped that whatever system is devised that it would be used for international events as well and these are the ones where checksum may be most required (shades of supposed collusion at Grandson IIRC). FWIW I am not bothered if it isn't checksum.

terrys wrote:
The 3-2-1-0 system isn't a checkpoint system and has plenty of support.

The tricky part of this one is getting the criteria for 1 and 2 right - see the ITC for details, as it isn't right now they way that is done and wasn't right befoe they changed it either IMO.

terrys wrote:
We could easily share 20 points between the players and give a 5pt bonus for breaking your opponent.
A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

Looks rather similar to the American DBM 15-0 scoring, which I personally like ...

I didn't make myself clear. I like the checksum system PLUS the bonus points. So yeah, my preference is:

A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

This would make players go for broke rather than stall for a win.

carlos wrote:I didn't make myself clear. I like the checksum system PLUS the bonus points. So yeah, my preference is:

A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

This would make players go for broke rather than stall for a win.

I would support this view for the same reason !
It is very close to the 3-2-1-0 system though...

Jerome

terrys wrote:We could easily share 20 points between the players and give a 5pt bonus for breaking your opponent.
A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

This is the best idea I have heard yet. I do not mind the lack of a checksum system, although this would have aspects of one. It would be easy to check back if any questions arise about the score for a game. And--best of all--this system gives an incentive to players who take the fight to the enemy.

Hurrah for 25-0!

Marc

I didn't make myself clear. I like the checksum system PLUS the bonus points. So yeah, my preference is:

A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

This would make players go for broke rather than stall for a win.

Same as I said before but better transcribed!
I totally agree with this system!

olivier wrote:

I didn't make myself clear. I like the checksum system PLUS the bonus points. So yeah, my preference is:

A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

This would make players go for broke rather than stall for a win.

Same as I said before but better transcribed!
I totally agree with this system!

But, of course, this system is not checksum and is open to collusion by dishonest players to score 15:15 instead of 10:10.

Of course we are all jolly nice chaps and would not do this, but we know it has happened in the past.

If both sides were close to breaking at the time limit, the temptation could be quite great for some people to boost their scores in this way. They could even rationalise it to themselves by saying "that is what would have happened if we had not been delayed by (the draw being late/my attack of diarrhoea/your having to go out for a cigarette every ten minutes - or other plausible reason for time unfairly lost from game time)".

A max win would be 25-0
A Draw would be 10-10
A mutual destruction would be 15-15

Mutual destruction must be considered as the best possible draw (10-10), not as a victory for both players.
A player can score a maximum of 9 points for a draw, other than mutual destruction.

Thinking about apply this score for a glicko rating system, this must be a zero-sum, progressive score.

Francisco.

Unfortunately scoring systems that are not checksum, have been abused in the past.

There is no point in asking players if the have a view on this because they simply dont think that this could happen. It does and it has and ist to late when it occusrs at an event. It just spolis the event for all concerned .

So it needs to be Checksum or close to it.

Addiitionally Checksum makes data entry a lot easier for organisersd and for something like Britcon this is a major consideration. If not checksum it increases data entry by 100%

Those who have been in the Britcon bunker on Friday night till 2 in the morning trying to find an error and then haviing to reenter will sympathise.

Regards
JDM