Slitherine

In the face of this apocalypse one wonders how the French, Gulf South or the ITC have managed all these years without a checksum system.

Regards,

Somebody who has thought about it and has been in the bunker ...

and some of thiose you list have been the victim of the problem

and next year you can do the data entry!

JDM

My bottom line is: players should never see hiding, stalling or running away as a way to win more tournament points.

Maybe only one player can score the extra points ? I remember playing last year in Seville where your won zero points for a draw and that seemed to encourage really aggressive play. Heck, I had skythians charging my early nubians !

I think that the risk of collusion between players to achieve a mutual-break draw, while not zero, is fairly low. The benefit to be gained from encouraging aggressive play is high. I think this strongly shifts the balance toward a non-zero sum scoring system for that reason.

However, the data entry point that JD brings up is a good one, especially for larger events. A zero sum system at least has the advantage of making it easier to spot typos and math errors coming from the players. At a small tournament it is fairly easy to check up on results that don't look right for some reason, or to go through the scoresheets to find which one you entered incorrectly. As the number of players ramps up the difficulty level increases dramatically, and a non-zero sum system adds to the hassle. Fair enough.

I still think that the balance weighs on the side of the non-zero system because of the potential to encourage the sort of play that makes everyone's tournament more fun. The score sheets can be designed to make it easy for the organizers to double-check the scores by listing number of BGs, number of BGs broken, etc., and to decide whether the win bonus has been correctly applied. If we suddenly see a rash of mutual destruction results I may have to eat my hat, but I do not think it likely.

That's my $.02,

Marc

From an organisers POV zero sum means you only have to enter one score per game and the system can work out the other one. Non zero sum means that you have to enter scores for each player. It might not sound like much but when for some reason the computer eats all three rounds of the Saturday data at a comp like Britcon with 200+ players only having to enter 300 scores to fix is is a lot better than having to enter 600

The none- checksum 25pt system works fine with a computerized system.

The scorer only needs to enter 2 things - the attrition points lost by each side - the program then calculates the points.
If one side has lost attrition points equal to his original army size, then his opponents get the bonus 5 points.
I think the misuse of 'mutual destructions' is a minor risk when weighed against the benefits of a system that encourages attacking play.

This would mean that the program would need to record the original army size for each player - at any time before entering the first score.
Since a nice-to-have would be a draw which included player name and army (to save players wasting time looking it up) - entering a players army/size at the start wouldn't be a big deal - since the information is already entered elsewhere.

A manual system would certainly be no more complicated than the current 32-0 system, and the method of calculating it would still be required for computerized comps, because most players will still wish to know how many points they got before seeing it on the scoresheet.

NB. One other benefit of a system like this (still applicable to the 32-0 system) is that the computer can retain fractions when calculating the scores, which will make it possible to seperate players on the same score to a much finer degree. I thought this over yesterday, and it would be possible to design a system that printed the scores to 2 decimal points (or more) if 2 or more player have the same (rounded) score.[/quote]

It gave Bob something to do

To keep data correction in perspective in my experience most errors come from players putting the down incorrectly on the score sheet which will cause problems for any scoring system - but the frequency this is done even in relatively simple systems is, IMO, a case for keeping the scoring system as simple as possible.

NB. One other benefit of a system like this (still applicable to the 32-0 system) is that the computer can retain fractions when calculating the scores, which will make it possible to seperate players on the same score to a much finer degree. I thought this over yesterday, and it would be possible to design a system that printed the scores to 2 decimal points (or more) if 2 or more player have the same (rounded) score

Beware, if you keep the decimal point, it favor big army vs small one.
neverless ,I agree with your non checksum 25 pts system.

Beware, if you keep the decimal point, it favor big army vs small one.

This isn't true ... After all losing 12 out of 16 is the same as 9 out of 12

I would round to the nearest whole number for each round of the competition - otherwise players my find they've lost a point because of the effect of adding 2 fractions - and only use the decimal point to seperate players on the same score.

Terry and I discussed the latest version today and my reservation about data entry on a non checksum system has been resolved, we have an idea that will fix this

so my concern about a checksum system is now at least partly sorted, we are getting there

Regards
JDM

If using a non-zero sum system, I would advocate including the ability to do something similar to the US 0-15 system, i.e. based on attrition but with double points for stuff you kill compared to stuff you lose. This shifts the cost/benefit balance towards doing damage even if you have no chance of breaking the opponent's army.

In a 25 point system this leads to, for example:

7 points for retaining own forces undamaged
14 points for damage done to the enemy
4 points for breaking the enemy

The formulas would be:

attrition ratio AR = attrition points lost/army size

score = 7 - 7*(own AR) + 14*(opponent AR) + (4: if opponent AR=1)

In fact, we can generalise this idea to play with the balance between killing, surviving and getting an army rout as follows:

a points for retaining own forces undamaged
b points for doing damage to the enemy
c points for causing an army rout

score = a -a*(own AR) + b*(opponent AR) + (c: if opponent AR=1)

Having looked at a few combinations, I think a=8, b=17, c=0 works better than the a=10, b=10, c=5 that Terry was suggesting. a=7, b=14, c=4 is also good if you want to reward costly army routs compared to large-margin winning draws.

If the computer already knows the army sizes then you only need to record and input the AP for the two players (2-list tournaments would need special handling). However, I have thought of a labour-saving scheme for the data entry person. Put an AR table (similar to the current one) on the score sheet, so players work out their AR from 0-10 (i.e. 10 times the AR as defined above). Replace the "10" with "A" (for "all"). Players then fill in their ARs on the sheet, which will be a single digit or letter. The data entry is then reduced to two characters, which is the same as at present. If players wish to work out their actual scores, they will usually need a calculator (as at present for BHGS DBM) but they could do without if the score system was e.g. a=10, b=20, c=0. This ditches the idea of 100 pts for a 4-game tournament, but none of the DBM systems had that, and no-one complained about it.

If you want to retain the ability to have a zero sum system that includes the win bonus (e.g. what we have been using up to now) you would need to have a=b and include another parameter in the formula:

d points for not being routed yourself (must equal c for zero-sum)

score = a -a*(own AR) + b*(opponent AR) + (c: if opponent AR=1) + (d: if own AR<1)

By basing the software around this formula, you allow tournament organisers a great deal of flexibility to choose a scoring system that suits them by setting the four parameters.

The current FOG system is a=13, b=13, c=3, d=3.
Standard DBM 0-10 is a=5, b=5, c=d=0.
US DBM 0-15 is a=5, b=10, c=d=0
For the continentals perhaps a=4, b=8, c=13, d=0 would be analogous to their 3-2-1-0 and a=5, b=5, c=100, d=0 to the +100 system.

By the way, Glicko will work with non-zero sum systems as it uses the share of the total points won by each player in the game. The absolute number of points doesn't matter.

My concern with this system is that it encourages a specific style of play. It induces players to use quick-kill armies and to ignore the slower winning armies - (like Arabs - with defensive spears and horse archers)

While I don't want to encourage defensive play, (and the game system doesn't do that). There are times when you come up against an army that your own army simply can't face head-on - given unfortunate terrain placement.
There should be a value given for a player making a resolute defence in a difficult situation.
One of the most enjoyable parts of the game is 'hanging on to the end' after you can see that your armies losing.
We certainly don't want to see players giving up simply because it looks like they're going to lose.

I prefer to see a player get his points for good generalship, not because of his opponents recklessness.


One other factor that should be considered is that it encourages players to artifically gererate a high casualty game in the interests of gaining a higher score. With the 7:14:4 system, if both sides were at 95% casualties, they would get 1:13:0 points - which would give them 14pts each. A battle with no casualties would give them 7pts each. So there is a possibility that 2 players could 'agree' on a result in order to score more points.
I know the vast majority of players wouldn't even consider this - but if you give people an incentive - someone, somewhere will take advantage of it.

terrys wrote:My concern with this system is that it encourages a specific style of play. It induces players to use quick-kill armies and to ignore the slower winning armies - (like Arabs - with defensive spears and horse archers)

10:10:5 favours quick kill armies as they have more chance of getting the 5. The slow-killers risk missing out on the 5.
9:16:0 for example gives slow kill armies an equal chance because they can still get most of the 16 even if they don't quite manage the army rout in the time available. For a symmetrical system this would be 12.5:12.5:0 or 12:12:1 if you want to stick to whole numbers and 25 pt hps.

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.

9:16:0 gives 9-9 for a bloodless draw, 16-16 for a mutual destruction and 25-0 for a max win.

While I don't want to encourage defensive play, (and the game system doesn't do that). There are times when you come up against an army that your own army simply can't face head-on - given unfortunate terrain placement.
There should be a value given for a player making a resolute defence in a difficult situation.
One of the most enjoyable parts of the game is 'hanging on to the end' after you can see that your armies losing.
We certainly don't want to see players giving up simply because it looks like they're going to lose.

I prefer to see a player get his points for good generalship, not because of his opponents recklessness.

A player can hang on to the end simply by playing very slowly.

One other factor that should be considered is that it encourages players to artifically gererate a high casualty game in the interests of gaining a higher score. With the 7:14:4 system, if both sides were at 95% casualties, they would get 1:13:0 points - which would give them 14pts each. A battle with no casualties would give them 7pts each. So there is a possibility that 2 players could 'agree' on a result in order to score more points.
I know the vast majority of players wouldn't even consider this - but if you give people an incentive - someone, somewhere will take advantage of it.

Has this happened in the USA under their DBM 0-15 system?


Different score systems incentivise various behaviours, depending on the motivations of the players, and opinions differ on what behaviours are desirable. In the end, giving organisers the option to use asymmetrical systems does not force them to use one. The Americans seem to like their system, and it makes sense to include them in your software, rather than have them develop their own methods as they did with DBM.

If a standard program is used to calculate results and draws having a parametersied system for score system means that the same program will be able to be used for multiple scoring systems and as a result it should make things easier for everyone.

I like the concept of a,b,c,d scoring options. Once we have a few more comps under out belts we can look at different scoring but still ues the same program.

One of the reasons for starting this post was so that we could get a universally agreed scoring system.
This was so that we could more easily compare results across different tournaments in different countries.

Having different scoring calculations available within the program is possibly - but will take longer to write.

I have noticed that the US players tend to be more 'Gung-Ho' in their approach to DBM, which (I think) leaves them at a disadvantage when coming up against the more considered approach of the Europeans. I'm begining to think that this is the effect of the scoring system(s) used in the US.

Fighting a battle is as much about keeping your own army intact as it is about destroying your opponent. Anything that destroys this balance leads to unrealistic army composition and tactics.
The US scoring system, and the 3:2:1 system were developed because it stopped players from sitting back and playing for a draw. This is already difficult to do in FoG even with the current system. I don't think we need more than a win bonus in order to incentivise players to attack.

terrys wrote:The US scoring system, and the 3:2:1 system were developed because it stopped players from sitting back and playing for a draw. This is already difficult to do in FoG even with the current system. I don't think we need more than a win bonus in order to incentivise players to attack.

I have to agree with Terry. We are not comparing like with like. Corner-sitting and table-sitting are simply not viable options in FoG (unlike in DBM) so it isn't necessary to try to prevent them by manipulating the scoring system - especially when such manipulation brings its own disadvantages, however minimal these may be thought to be.

Also, it may be worth restating, for the benefit of those uninterested in tournaments, that the above discussion is only about tournament scoring systems. The victory conditions in the rules are fine for non-tournament play.

We would not want anyone approaching the rules afresh to think that they need a computer program to work out the result of their games!

terrys wrote:One of the reasons for starting this post was so that we could get a universally agreed scoring system.
This was so that we could more easily compare results across different tournaments in different countries.

Having different scoring calculations available within the program is possibly - but will take longer to write.

I have noticed that the US players tend to be more 'Gung-Ho' in their approach to DBM, which (I think) leaves them at a disadvantage when coming up against the more considered approach of the Europeans. I'm begining to think that this is the effect of the scoring system(s) used in the US.

This is not only a US but also a continental european + irish approach ! i.e. non-british

terrys wrote:Fighting a battle is as much about keeping your own army intact as it is about destroying your opponent. Anything that destroys this balance leads to unrealistic army composition and tactics.
The US scoring system, and the 3:2:1 system were developed because it stopped players from sitting back and playing for a draw. This is already difficult to do in FoG even with the current system. I don't think we need more than a win bonus in order to incentivise players to attack.

Any scoring system allowing a player to win a tournament without winning is a non-sense.
The 310 system would mean that a victory must always be better than 2 draws which is a good basis for any thought on a scoring system (to my point of view and not only).

Jerome

Pikeaddict wrote:Any scoring system allowing a player to win a tournament without winning is a non-sense.
The 310 system would mean that a victory must always be better than 2 draws which is a good basis for any thought on a scoring system (to my point of view and not only).

Jerome

I am not convinced.

In Port Elizabeth for example had we played with 321 I would have finished second with four wins and two "draws" behind Corbon with five wins and a defeat. The problem is that Corbon's defeat was at my hands and meant that his 5th game was easier as a result.

I played my last four games on table 1 or 2 and my last three on top table. Corbon dropped a bit when I beat him then came back to place 2nd.

I don't object to 321 scoring in tournaments with fewer games but in a big comp it is silly for a player to have a record of 4 wins a narrow winning draw and a draw yet finish behind a player with five wins and a defeat to the player that only had the four wins.

Hammy

hammy wrote:I don't object to 321 scoring in tournaments with fewer games but in a big comp it is silly for a player to have a record of 4 wins a narrow winning draw and a draw yet finish behind a player with five wins and a defeat to the player that only had the four wins.

What this says to me is that all attempts to manipulate player behaviour by means of scoring systems will cause anomalies.

I submit that with FoG there is no need to manipulate player behaviour by such crude and heavy-handed means because the types of behaviour that were a problem in DBM and that these systems were intended to address are not a problem in FoG.

Moroever, the more the scoring system favours outright wins, the more unfair disadvantage there is to the player whose opponent deliberately plays slowly to avoid an army break. Also the more advantage that (unfairly) accrues to players drawn against ill-matched opponents who can easily be defeated in the time limit.

There is in fact nothing wrong with the system that has been used in the last couple of FoG tournaments in UK and I would support a variant of this. I am completely against any system that heavily penalises incomplete games.

rbodleyscott wrote:
I submit that with FoG there is no need to manipulate player behaviour by such crude and heavy-handed means because the types of behaviour that were a problem in DBM and that these systems were intended to address are not a problem in FoG.

This is a very good point, which implies a symmetrical system should at least tried out first.

Pikeaddict wrote:
Any scoring system allowing a player to win a tournament without winning is a non-sense.
The 310 system would mean that a victory must always be better than 2 draws which is a good basis for any thought on a scoring system (to my point of view and not only).

Jerome

Moroever, the more the scoring system favours outright wins, the more unfair disadvantage there is to the player whose opponent deliberately plays slowly to avoid an army break. Also the more advantage that (unfairly) accrues to players drawn against ill-matched opponents who can easily be defeated in the time limit.

Jerome and RBS are in complete disagreement here.

I'd like players to consider the question:

Who deserves more points out of

a player who causes 90% attrition (i.e. not quite an army rout) on the opponent and loses nothing
or
a player who causes 100% attrition (army rout) on his opponent and suffers 90% attrition himself ?

Can we set up a poll?

There is in fact nothing wrong with the system that has been used in the last couple of FoG tournaments in UK and I would support a variant of this. I am completely against any system that heavily penalises incomplete games.

Based on the above, I would suggest 12:12:1 for RBS and 8:8:9 for Jerome.