Slitherine

If there are a greater spread of wins and losses then I accept that there may be a greater number of people on the same points. This is where the table of points above will help.

And don't forget the extra potential for differentation based on Bp that is a part of that scoring.
The more I think of my game vs pantherboy the sillier it gets. One loss and one draw for me but the points are 33/27 in my favour.
I actually played kamikaze style with the Parthians. Knowing I would get nothing for a draw meant I just charged forward and tried to inflict as many Bp as possible.
Not the way one usually play.
PS-Buy some expansions man! At least IF

People

If you look at the table I suggested you would get a slightly bigger spread with the bonus point, and might satisfy the desire to award big wins.
In a mirror tourney - two games a round

10 pts. - two big wins
9 pts. - two wins - 1 big
8 pts. - two wins
7 pts. - 1 big win, draw
6 pts. - 1 win 1 draw, OR big win, narrow loss
5 pts. - 1 win 1 narrow loss, OR 1 big win 1 loss
4 pts. - 2 draws
3 pts. - 1 draw 1 narrow loss
2 pts. - 2 narrow losses OR draw and loss
1 pts. - 1 narrow loss 1 loss

I would also award a bonus point for a score draw, as opposed to a turn expired draw.

The advantage of this is there is an incentive to play to the end. Either to score a big win - or to kill some of the enemy and gain a narrow loss. Which has some historical basis as well. Wanting to smash the enemy to prevent his return - or inflict damage on the victor as he is likely to have to fight again, before he is ready.

One of the main reasons I dislike whether a win is big or not is due to context. In tournaments, leagues etc. where you are playing isolated battles then the final result of a win or a loss is the only important factor. This helps alleviate imbalances between armies, players, luck throughout the battle etc. In a strategic campaign where your army must fight continuosly, issues of supply are relevant, alliances etc. then the extent of the win is very important and how that impacts your capability to continue your nation's interests. Thus we can get situations where a player loses but wins because because he inflicts such losses in the battle that are recoverable for him but devastating to the opponent (ala Pyrrhus and his Pyrrhic victory). Context is important here and as such the simple nature of whether you won or loss becomes irrelevant in the greater scheme of things. Also imbalances in armies, players etc. matters not as new weapons are available to combat them e.g. alliances, economics etc.

I also see no difficulties arrising with a 100 players. As I said give me the raw data or post it here and then see what people come up with.

Also I feel the two matches should not be independant of each other (as you have it now) as that creates issues where imbalances exist. The purpose of a mirror match is to find the better player of that particular scenario by taking into account both sides otherwise you need to have equal DAG armies on the same map for both battles. This also leads to the purpose of having a scoring system that will award a win to the player who performs best overall and fairly rank those in descending order after.

Sigh...

When I looked at the stats there were no byes in the first 3 placed players. I look at stats all the time and know exactly how to interpret them or I would not be able to do my job or have won every competition going related to wargaming all around the world. I have no problem changing the score allocated to a bye but this is nothing to do with the basic design of the system.

Please give us some credit.

There is no raw data on the results, only the scores.

The aim is not to get the best player at the top of each competition. The system is designed after 20 years of experience of running tournaments and scoring them. I'm not saying its perfect but please accept that we will be having a system that rewards big wins over small wins as this is a core design feature which works well from experience.

There may be some bugs in how scoring is calculated that mean it is not working as intended but I do not want a win/lose/draw system. It does not work with so many players over so few rounds. Give an example of any other system which uses this type of scoring with so many players. E.g. sports only use win/lose/draw scoring when there are very few teams (e.g. Champions League) or you get teh chance to play everyone (e.g. Premiership).

If you're not willing to work within the criteria I need then I'll have to spend some time working on this later when I have time.

Well the tournie winner hidde had a bye first round and his opponent in the second round threw in the towel towards the end even though he could of taken a few more points of his score (based off what hidde said to me) and then outscored me when he drew and lost. I hate to keep using him as an example but he has a trifecta of situations which show the weaknesses in your system. After second and third the next six players have byes and one has two. I don't know what table your looking at.

There is no raw data so why would I give you any credit. That is one of the first things you'd want to be accumulating if you wanted to accurately analyze a system. But it is possible you have misunderstood what the request entailed so in simpler terms I just need the actual score for each battle e.g. Romans 45/45 Germans 30/45 with the players identified. I would be truly suprised if you cannot gather this information as you stated to me in previous mails that you were able to watch replays of matches etc. That would indicate that if you spent thirty minutes (about what I think would be required) to bring up each match, without the need to replay, and then record the game totals.

So I take it your aim is not to have the player who performed best in the tournie win. Well I can't help you here as that defies the purpose of tournaments. From my experience the goal is to find the best player/team in the tournament. If your worried that a minority of players will dominate then create tournies based on rankings which can be taken from past performances in tournies. Awfully simple to achieve and implement without the need of making it overly complex.

I think I've said this before but just because something has been used for 20 years doesn't immediately make it good. That is why innovation exists.

If you want the help from the community then don't wave the flag of how good you are but instead delve into the issues at hand and provide the information requested. Then we can do something worthwhile rather than belabor this point. If you want big wins rewarded though you don't provide any reasoning to why it is such a good idea then we can use that as a guideline when attempting to alter the system. If your just dogmatic and in love with your own ideas then just say so and keep what you have and then people will stop wasting time reasoning with you.

For my 2 cents, I am like the rest of the guys a little confused by the scoring system. Some results make sense, but others I just cannot work out at all. In particular the Hidde v Pantherboy game. How can Hidde lose one battle and draw the other battle yet score 33pts to PB's 27pts (with no time outs on either side apparently) ? Surely a bug?

As for players getting an advantage when drawn against less experienced foes or vise versa, I like it, it is just the luck of the draw similar to the English football FA Cup. Sometimes you have to play Chelsea, Arsenal and Man U to get to the final, sometimes you get to play Plymouth, Halifax and Kettering Town. If this happens because of byes, its just tough (or good) luck. The advantage of the current system is that we won't see the same old people winning every tournament. We will also get to play a wider variety of other people/skill levels instead of the same ones we see in the league games. After all, tournaments are just a bit of fun. The real proof of competency will always be the league games.

Steve, I am happy to supply my 'raw stats' for you to number crunch. Finding flaws and bugs is important. I'm sure others will be happy to supply their game stats too while the numbers are still in the Results inbox.

Cheers ianiow. Thanks for your opinion. I personally have no trouble with the drawing of players and enjoy the greater variety fielded in tournies. I just prefer that players who are fortunate to bye or get an easy match-up recieve no more benefit than an easy round. But if people are willing to contibute real results then we may be able to discover a better solution. I am considering just doing a mock-up 100 player tournie with fictional results that would include all the situations encountered as an alternative but it would be nice not to have to do so much work.

My two cents is that the scoring system is ok, but there are a few small issues.

1) Timed victories give too many points to the winning side. Perhaps the points for timed victories should be capped at the mean of the untimed victories for that particular round. This would prevent someone from obtaining a 60 point score because there opponent was a no show.

2) If someone has a true "bye" in a round they should be awarded 30 points not 60.

3) I'm not sure about this but it seems the way the score is calculated it does not take into account any additional losses inflicted on the loser above the trigger score. I.e. if I win 52/44 to 30/44, the extra 6 units I killed are not used in the score calculation. Seems like they should be.

Reading the new manual has opened up new possibilities regarding scoring. I post the relevant section from the manual.

Decisive Victory
When, at the end of any player turn, an opponent's total break points equal or exceed their army score the game is over and you are victorious!
If at the end of any player turn your total break points equal or exceed your army score the game is over and you have been defeated!
Marginal Victory
If the game continues to the end of the last player turn without a Decisive Victory, then the two sides divide their total break points by their army score. If the difference in the two army scores is equal to or greater than 25% then the player with the lowest break point is declared the winner with Marginal Victory.
Draw
If, at the end of any player turn, both sides total break points equal or exceed their army score the game is over, both sides have exhausted their ability to continue the battle is a draw!
If the game continues to the end of the last player turn and the game has not ended for any of the above reasons then neither side has achieved a satisfactory margin of victory over the other the battle is a draw!

So now more points can be awarded for Decisive victories rather than marginal victories. Surely this would meet Iain's need to reward BIG wins without any complicated maths related to BPs?

Now, my only concern is that the results posting doesn't always seem to reflect this logic, although I've read a few posts recently that state the results window does reflect this. Certainly in most (all?) the informal competitions that we've been running for the last year or so have treated the Marginal Victory result as a draw.

If we tabularise the possible combinations of combinations of Decisive victory, Marginal Victory, Draw, Marginal Loss, Decisive Loss plus the BP weightings then there are plenty more points that can be awarded for each paired game making the chance of lots of similar scores at the end of 5 rounds being reduced. e.g.
- 2x Decisive Victory
- Decisive and Marginal Victory
- 2x Marginal Victory
- Decisive Victory and Draw
- Marginal Victory and Draw
- Decisive Victory and Decisive Loss (net positive BPs), or Marginal Victory and Marginal Loss (net positive BPs)
- 2 x Draw, or Decisive Victory and Decisive Loss (net equal BPs), or Marginal Victory and Marginal Loss (net equal BPs)
- Decisive Victory and Decisive Loss (net negative BPs), or Marginal Victory and Marginal Loss (net negative BPs)
- Marginal Loss and Draw
- Decisive Loss and Draw
- 2x Marginal Loss
- Decisive and Marginal Loss
- 2x Decisive Loss

So, if 12 points were awarded to the 2x Decisive Victory and reducing by 1 to 0 points for 2x Decisive Loss, then this should give a reasonable spread. Another advantage is that it is easy to understand.

For non-played matches then I'd award 2x Marginal Victory (defaulted) or 2x Marginal Loss (defaulter)... but we could argue the pros and cons of this... and I'm sure we will!

The one thing I'd ask is that the results shown outside the results table should reflect this too!

Hmm not sure I like the marginal victory thing here

example (if i undersatnd correctly)
player A leads a 50 bp army
player B a 40 bp army
i dont know if this is an "average" BP size diffence between armies but seems reasonable

A wins 38/50 to 40/40

that is a 24% differential and thus a marginal victory for A , if it was 37/50 though , then it would be 26%, a Major Victory. 1 bp makes a huge difference . Should it?

Im a big fan of the concept that , like sports and likly battles , the final outcome is what counts , you either win or lose, in the end noone cares you missed ist place by a *** hair

Counting scores etc has some merit but in the end, i think it encourages gamemanship rather than tactics and could result in people supressing the worthy goal of trying to beat the enemy despite the odds, and engage in bean counting. A second battle in a mirror match will play out so much differntly if you recognize "I just need to take no higher than 30 bps' and inflict more than 10 and I will come out ahead" , vs "I lost the ist favoured battle, now i really need to be a tactical genious to WIN the non favoured battle".


My likly less than 2 cents worth of opinion.

TheGrayMouser wrote:Hmm not sure I like the marginal victory thing here

example (if i undersatnd correctly)
player A leads a 50 bp army
player B a 40 bp army
i dont know if this is an "average" BP size diffence between armies but seems reasonable

A wins 38/50 to 40/40

that is a 24% differential and thus a marginal victory for A , if it was 37/50 though , then it would be 26%, a Major Victory. 1 bp makes a huge difference . Should it?

Im a big fan of the concept that , like sports and likly battles , the final outcome is what counts , you either win or lose, in the end noone cares you missed ist place by a *** hair

Counting scores etc has some merit but in the end, i think it encourages gamemanship rather than tactics and could result in people supressing the worthy goal of trying to beat the enemy despite the odds, and engage in bean counting. A second battle in a mirror match will play out so much differntly if you recognize "I just need to take no higher than 30 bps' and inflict more than 10 and I will come out ahead" , vs "I lost the ist favoured battle, now i really need to be a tactical genious to WIN the non favoured battle".


My likly less than 2 cents worth of opinion.

With any sort of victory other than 'a win' there will always be a cut-off point, so at that point the 1point will make all the difference.... well, 1 point in the scoring system proposed.

And while I agree 'a win's a win', many games have degrees of victory ,e.g. heroic, decisive, marginal etc. so it's not that unusual. Mind you, Pyrrhus would probably argue that not all win's are equal

I've always found the concept of degrees of win in games purely a way of massaging ones ego. As I stated before unless your incorporating a strategic level then the only relevant factor is the win or not. As soon as you grade the quality it becomes meaningless as it has no capacity to take into regard factors such as luck, skill, experience with the game/scenario, terrain selection etc. And due to these factors awarding points based on such an outcome creates issues regarding scoring for draws, playing imbalanced scenarios etc.

I know this will be stating the obvious but

1 A new tournament should be promoted a couple of weeks before it begins, showing all of the dates and so ensuring that players are totally committed to playing.

2 Increase the time period to finish each round just in case some unforseen circumstance prevents a player from playing 1 or 2 days.

3 Scoring system. I like the idea of a larger points system? That is 21 vs 9, 18 vs 12 etc, rather than a straightforward 3 for a win, 1 for a draw, 0 for a loss but that system has to be tried and tested so it is in perfect working order.

4 Swiss system, making sure the same players do not play again. This is a must for me.

If 3 and 4 can be sorted I will definitely play.

Again would like a proper competition, that is 6 games using an army of your choosing based on a specific time period and just the one game. It could be from 100BC to 100 AD and include all of the armies from that time period as the choices for players. Would probably include IF, ROR and LT.

I don't mind historical scenarios but I'd prefer an "open" competition.

iainmcneil wrote: There may be some bugs in how scoring is calculated that mean it is not working as intended but I do not want a win/lose/draw system. It does not work with so many players over so few rounds. Give an example of any other system which uses this type of scoring with so many players. E.g. sports only use win/lose/draw scoring when there are very few teams (e.g. Champions League) or you get teh chance to play everyone (e.g. Premiership).

Iain

I have no particular axe to grind here and I like your scoring system. My 2p worth is it is sound though that will always be debatable for all the reasons others have already discussed.

However I do think the byes and time-outs with no bp loss scoring 0+0 vs 30+30 are putting a skew on final scores that is the root cause of this discussion.

I have a potential solution that you may or may not have considered already.

I have just run it on the LT tornament scores and to my neutral eye it looks to have provided a fair and reasonable out come.

The rationale is this
Change none of your scoring system except this - The final 'Ranking Score' is a moderated score of points per game played calculated to 2dp
Byes and no bp loss time-outs are discounted as games played unless you are the one timed out in which case the penalty is tough.

Where a player is timed out with no bp loss that game counts as played for the player timed out - who also scores zero point. Effect on the ranking score is it goes down; by 1/4 in a 4 round tournament.
The player not timed out receives no score but does not count a game played - effect on the Ranking Score is none, it is simply calculated on the other played games.

Byes work on the same basis.

I would suggest that this is simple, workable and does not change any of your tournament principles or fundamental scoring particularly if the ranking score goes to 2dp.

It took me 45 minutes to apply this to the LT results this morning and the number of games affected were small, however it does alter positions in the ranking - considerably in some cases.
Ianiow is the winner (just) in this alternative scoring - Not that that means anything at all - sorry ianiow .

Here is a link to the result showing the adjustments
https://docs.google.com/viewer?a=v&pid= ... OTY5&hl=en

Hope you do not mind the suggestion - should be good for a bit more discussion

Triarius wrote:
It took me 45 minutes to apply this to the LT results this morning and the number of games affected were small, however it does alter positions in the ranking - considerably in some cases.
Ianiow is the winner (just) in this alternative scoring

I like your system Triarius

Triarius wrote:

iainmcneil wrote: There may be some bugs in how scoring is calculated that mean it is not working as intended but I do not want a win/lose/draw system. It does not work with so many players over so few rounds. Give an example of any other system which uses this type of scoring with so many players. E.g. sports only use win/lose/draw scoring when there are very few teams (e.g. Champions League) or you get teh chance to play everyone (e.g. Premiership).

Iain

I have no particular axe to grind here and I like your scoring system. My 2p worth is it is sound though that will always be debatable for all the reasons others have already discussed.

However I do think the byes and time-outs with no bp loss scoring 0+0 vs 30+30 are putting a skew on final scores that is the root cause of this discussion.

I have a potential solution that you may or may not have considered already.

I have just run it on the LT tornament scores and to my neutral eye it looks to have provided a fair and reasonable out come.

The rationale is this
Change none of your scoring system except this - The final 'Ranking Score' is a moderated score of points per game played calculated to 2dp
Byes and no bp loss time-outs are discounted as games played unless you are the one timed out in which case the penalty is tough.

Where a player is timed out with no bp loss that game counts as played for the player timed out - who also scores zero point. Effect on the ranking score is it goes down; by 1/4 in a 4 round tournament.
The player not timed out receives no score but does not count a game played - effect on the Ranking Score is none, it is simply calculated on the other played games.

Byes work on the same basis.

I would suggest that this is simple, workable and does not change any of your tournament principles or fundamental scoring particularly if the ranking score goes to 2dp.

It took me 45 minutes to apply this to the LT results this morning and the number of games affected were small, however it does alter positions in the ranking - considerably in some cases.
Ianiow is the winner (just) in this alternative scoring - Not that that means anything at all - sorry ianiow .

Here is a link to the result showing the adjustments
https://docs.google.com/viewer?a=v&pid= ... OTY5&hl=en

Hope you do not mind the suggestion - should be good for a bit more discussion

I don't know if it is deliberate, but I can't view the Google Doc.

Hi Morbio
I had set the file share properties to 'anyone with link'.
Are others trying and failing to open link?

I can't link to it either

pantherboy wrote:I can't link to it either

ditto