• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Hiliadan

Major
Jun 17, 2018
614
6
A poll was run in April-May 2018 to determine what type of tournament people wanted to play: https://www.the-battlefield.com/aow3/index.php?page=bf_poll&pollnumber=32 A 2vs2 format got the most votes so a team tournaemnt was organized: https://www.the-battlefield.com/aow3/index.php?page=tourneymain&tourneyid=14

A Duel Tournament got the 2nd most votes so it's now time to start organizing this Duel Tournament!

In the continuation of this thread on the old forum http://aow.triumph.net/forums/topic/preparation-of-next-pbem-tournament, I'd like to open discussions about the settings of that tournament.
Personally, I'd keep all the settings and rules from the 2vs2 Tournament except:
- map size = Tournament Square
- possibly Fast Game Speed (to be tested) to boost production classes vs mind-control strategies
- add rules about no game restart even if bad conditions


What is everyone's opinion?
 
Have created a game with "FAST". But using small map for 2on2, and no one like to play with small Map!

Posted another game with alternative Settings, but it shall be settings for non-Veteran and new players. Because the games are very frustrating played by elite.

Think this is one Point why only top Players are active on Battlefield (except admins and friends). The Frustration factor is great when you play against top players. (Cat and Mouse Games, Destroying Cities, rewards with sites - spamming elemtals and t4 Monsters, extremly mind Control , economical overpowering). But Nothing to do with Tournament settings Just another view. You see when someone post something like Middle Earth (and please no Veteran), Players will join. but not for normal ladder games.
 
i dont have any experience wiht fast speed either. hence im a bit averse to it... (But the problem still subsists, the conversion classes are too strong. mc-abilites should be all reduced in strength by 1 or 2.)

i would like to propose to play with the setting: "random heros match race". Why? Yes, there is a good reason to play without it. But for races which have a terrain mp bonus (tigran, frostling, dwarves, goblins, elves, haflings, so all but humans and draconians), it is really a source of irritation and disruption, especially in the beginning, if one hero makes the whole stack slower. Not only this, but some heros will have a terrain penalty on the wrong terrain. Which leads to fumbled attacks and is especially bad for hafling heros. this is by now enough reason for me to prefer to play with this setting on. Some heros are stronger than others and we gotta tackle that too. But I d like to play with "random heros match race" setting on in the tournament. phirpo
 
Fast speed does not make consensus. Some players claim it helps Rogue and Theo because they get faster access to Bards and Evangelists (obviously they do, that's the whole point of making research faster), but without any real analysis of the situation. My limited experience so far (2 games) is that production classes benefit as much as Rogue and Theo and a rapid calculation shows that Bards are acquired only 2-3 turns earlier than with Normal speed, while Evangelists are maybe aquired 4-5 turns earlier, which should not make a big difference (for Rogue at least).

Anyway, I'm thinking about another change to the settings (probably not for a tournament as I don't plan to organize any, but for "standard games"): removing Treasures (gold and mana stashes). That should hurt more Rogue and Theo who can easily pick them with Crows and Cherubs. It should also make it harder to sustain mind-controlled units as less gold will be available. And it should further slow down the pace of the game, advantaging production strategies.
Thoughts?
 
Maybe the both classed shall be banned except for finals. If both can use this class it is not a disadvantage for the match, when banned for all, it is fair.
It shall be known which class are stronger and weaker. But i fear the same discussion with other classes....after such banning.

Another one I tried out in alternative settings. You can only use converted units against indies, not against humans.
That was not a compulsion to dissolve units, but you could not use them against your Opponent (to rush your Opponent). A cool idea can be: When converted Units can be buy with Gold after battle.....maybe in the next inn

Further in this settings
Destroying and creating cities was also banned, so it was no longer possible to support huge armies due to a good economy.
Furthermore, a component Visible treasury was to adjust to few or none.

But make the rules a bit complex
 
Hm, strange. I thought I aswered to this topic already but it seems like I forgot to click on the button to send it. Banning classes is imo absolutely no solution. Rogue is strong but far away from OP. If you ban theo and rogue there will be other classes filling this postion. AD and Necro are coming to my mind and banning classes would change the whole game.

Removing treasures is a good idea. It hurts all classes/races but obviously it hurts the classes with very movable scouts with good sight more. Theo and rogue have the best scouts in the game so it would definetly hurt them more than other classes. Seems like a plan for me! :)
 
Last edited:
  • 1Like
Reactions:
i think there is much possibilites. But fair games are not welcome. Use Mirror map and Mirror Leaders...but i think no one will like it.

I'm a fan of complete random games, get place in world and make the best. not pre-select race or class, not pre-select map, Maybe in scaling from small to large with or w/o ug.
But Maybe not complete, going too Long.

An idea can be, the admin select the leader not the Player and use randoms.
 
Regarding Fast Speed, I changed my mind: the issue is not really with "normal" research, I still think it does not boost Rogue or Theocrat through Bards or Evangelists that much.
But -25% RP cost means all the sites giving RP as a reward are much more useful and that's an issue.

So I'd rather go with either +33% RP per building (Lab, Observatory...) or remove the RP change altogether and just have a Fast Population settings which keeps only the population change.

A "championship" is being considered. Each group has 4 players who play each other exactly once during a season (so about 6-9 months per season probably). Then the #1 is promoted to the league above and the #3 and #4 are relegated to the league below (only the #2 stays within his current league). With 2 or 3 league levels (so 12 to 28 players).
There would be no time limit on matches but if a match last longer than 2 months, players should signal it (both are penalised with less points if the match does not end) and judges would see what to do. It would be possible to play all 3 matches in parallel if players want.

The fact that only 3 matches mean it's different for the class/race pick restrictions. I guess at least you should not be able to pick the same class or race during a season. But that's less restrictive than during a tournament. So maybe only 3 classes and 3 races randomly chosen should be allowed each season. To be discussed.
 
The fact that only 3 matches mean it's different for the class/race pick restrictions. I guess at least you should not be able to pick the same class or race during a season. But that's less restrictive than during a tournament. So maybe only 3 classes and 3 races randomly chosen should be allowed each season. To be discussed.
Maybe just vetoing classes/races every game by players one by one and pick any from the rest?
It is simple and do not required to track choices during the season.
 
Maybe just vetoing classes/races every game by players one by one and pick any from the rest?


I'm unsure what you mean? Do you mean the same system as in current tournament? So if in season 1 you played 3 classes and in season 2, you played 3 more. In season 3 you need to start with the last one (there are 7 classes) you haven't played yet?
 
I'm unsure what you mean? Do you mean the same system as in current tournament? So if in season 1 you played 3 classes and in season 2, you played 3 more. In season 3 you need to start with the last one (there are 7 classes) you haven't played yet?
Before each game players veto/ban classes from the entire pool.

Player1 bans ArchDruid.
Player2 bans Theocrat.
Player1 bans Dreadnought.
Player2 bans Warlord.

Sorcerer, Necromancer, Rogue to pick.

Player1 picks Necormancer.
Player2 picks Rogue.

Any variations can be applied such as allowing mirrors (so second player picks from 3 classes as well), ban some classes randomly before voting.
The same for races.
 
So you guys are looking for a middle way between total random and total freedom in choosing your class/race combination (and specialization?).

There's a balance mod out but you still need a way to play with your most preferred or to avoid playing against your most feared combination?
And you propose to add another card-like layer of strategy before actual game begins?
:D

Why not the following, then? (a weighted random draw)
(I had a couple hours to spend, anyways)

1. Principles:

- Players have a set number of 'wish' points to be allocated to chosen race(s) and/or class(es) in order to affect randomness in combinations, as compared to a pure random determination. They are each free to spend or not any number of these points.
- If a player doesn't use any point, the combination is purely random.
- If a player use points for preferred race(s) and/or class(es), the odds these races and/or classes are obtained are higher, except if another player chooses the same preferred race(s) and/or class(es).
- This way, a player having rather despised/overlooked preferred combinations would have a greater chance to obtain them than when pure random is used.
- This way, a player having rather popular preferred combinations would have less chance to obtain them, but no less than when pure random is used.
- Allocating points has two effects: changing the odds between the two pools of preferred races/classes and non-preferred races/classes (for each player), and changing (weighting) the odds between races/classes in the preferred pool (if not, they would be equiprobable). The later also depends on the other players' choices, see point 2 thereafter.
- The idea is to affect a random draw for each player, while keeping it non arcane, so that everybody can easily know the odds for each class and race.


(any single figure below can be adapted to what suits the community most)

  • Each player gets 4 points to favor preferred races and 3 points to favor classes.
  • No more than 2 points can be put in any given race or class. Any one gets either 0, 1, or 2 points by each player.
  • This means that each player can favor up to 4 races out of 9, and up to 3 classes out of 7, so a little less than 50% for both.
  • Possible combinations of point allocation: from 0/0/0/0 to 1/1/1/1 to 2/2/0/0 preferred races, and from 0/0/0 to 1/1/1 to 2/1/0 preferred classes (all 0 means that pure random is used for that player).
  • Actual weighting is yet to be refined, but as for classes, let's only consider those three allocations: 0/0/0, 1/0/0, and 2/0/0
    0/0/0 means that each class gets a 1/7 chance to be drawn (~14%) by this player.
    1/0/0 means that the preferred class gets a 2/7 chance to be drawn (~28%), while the remaining classes share the complementary odds (5/7 or 72%, so 5/7*1/6 or 12% each).
    2/0/0 means that the preferred class gets a 3/7 chance to be drawn (~43%), while the remaining classes share a 4/7 or 57% odds (so 4/7*1/6 or 9.5% each).
  • Case of multiple choice by one player:
    If a player elects to favor several classes or races (e.g. putting in 1/1/0 points in classes), the summed up number of points allocated amounts to between 2 and the max number (3 for classes), and this means (for classes) that the preferred pool either gets 2/7, 3/7 or 4/7 (~57%) chances to be drawn. Then a second draw is ran. If preferred pool was drawn, then classes in this pool are weighted according to allocated points (1/1/0 means 50%/50%, 2/1/0 means 66%/33%, and 1/1/1 means 33%/33%/33%). If it was not, equal odds are used to draw the class in the non-preferred pool (with a 20%, a 20%, or a 25% chance for each class in that pool, respectively). This leads to individual chances, in the 2/1/0 choice, of: 4/7*2/3 (~38%), 4/7*1/3 (~19%), and 3/7*1/5 (~8.6%) for each last 5 classes. (total of 38+19+5*9=~100)
    Another solution is to keep individual weights as they are defined in previous point, so that a 2/1/0 choice leads to 2/7 and 3/7 chances for two preferred classes (a total of 59% chance to draw one of those two classes, instead of 57% above), and then the other 5 classes each get a 2/7 chance divided by 5 (~5.4%). (total of 43+28+5*6=~100)
  • Case of popular class(es)/race(s):
    In case several players allocate points in the same class(es)/race(s), then instead of letting do and drawing combinations independently for each player, as in previous point, a secondary weighting procedure is used to discourage the trend for a few, "OP" combinations, and bolster the variety, should this be wanted as said in the principles. So their choice is degraded and the situation gets closer to pure random. Starting with the first proposal in previous pointThe idea is that the bonus for choosing a preferred class/race is degraded by the bid of a player proportional to the total number of allocated points in this class/race. The chance of drawing a peculiar class/race is now: (1 + (points allocated by player)^2/(total number of allocated points by players))/(number of available class/race), because in this proposal, the allocated points (1 or 2) equal the bonus (1/7 or 2/7 for classes).
    For example, if two players put 1 point in the same class, instead of each having a 2/7 chance (~28%) to draw this class, they now both get a 1.5/7 chance (1+1*1/2 out of 7, or ~21%). The same degradation would happen if both put 2 points in the same class (because 2/4 =1/2), but this would apply to a greater added chance of 2/7 instead of 1/7, for a resulting chance of (1+2*2/4 out of 7, or ~28%).
    Another example: Four players allocate points in the same class (2, 2, 1, and 1 points, respectively). Players who invested one point and should have had a 2/7 chance (~28%) now gets a (1+1*1/6 = 1.17)/7 chance (~17%), while both players who should have had a 3/7 chance (~43%) now get a (1+2*2/6 = 1.67)/7 chance (~24%). Still better than the default 1/7 chance, but clearly not as high as expected, because this class is too much popular and that's against the variety in the competition (should this rule need to be enforced).
    Full cases for three players choosing the same class (percents are chance to draw that class):
    0 0 0 -> 14% 14% 14%
    1 0 0 -> 29% 14% 14%
    1 1 0 -> 21% 21% 14%
    1 1 1 -> 19% 19% 19%
    2 0 0 -> 43% 14% 14%
    2 1 0 -> 33% 24% 14%
    2 1 1 -> 29% 21% 21%
    2 2 0 -> 29% 29% 14%
    2 2 1 -> 26% 26% 20%
    2 2 2 -> 24% 24% 24%
    In this example, investing 1 point on a class warrants a chance of 19%-29% instead of default 14%, and investing 2 points warrants a chance of 24%-43%.
    Another formula (remove the square factor) would lead to a greater degradation.
    Another approach could add the number of players into the formula, so that two players choosing the same class among 4 players have better chances than when they are in a 1v1 game (for the same invested points).
  • I didn't check what happens and how to articulate both algorithms when players select several same classes, but the general ideas should stand.
 
I'm still unconvinced by keika and Rodmar's proposals. The rules should be simple and not rely too much on players discussing between each other, because my experience is that it's going to take ages or fail. So it's better if judges decide for players and it can be communicated very clearly and be something applicable to ALL players for the season.
So I'd rather pick 3 classes and 3 races (or maybe rather 4, so that even at your last match, you still have a choice between two) that everyone can play. The drawback, as pointed by keika, is that judges still need to track the choices of players over the season, and it's time consuming for judges. I'm going to see with Mark if we can set up an automatic tracking system, that would REALLY help judges (even with a different system than the one I suggest).

Meanwhile, here are the rules and settings I suggest (changes from 2019 Duel Tournament in red):


Settings

Balance mod: http://steamcommunity.com/sharedfiles/filedetails/?id=661597466
Strong Defender Normal Reward mod: https://steamcommunity.com/sharedfiles/filedetails/?id=1374641017
Later seasons will include the Shadow Realm expansion.

Turn timer = 1 day (activate only after the 1st turn is played)
Small Square Map with Surface and Underground on Continents
Tournament Game flow

Start town: Village
Starting units: Weak
Starting distance: Far
Roads: Few
Roaming units: Average
Treasures: None
Cities: Few
Dwellings: Few
Ressource structures: Average
Visit structure: Average
Treasure structures: Average

Water Slider at 25%
Undiggable walls and Diggable Wall Sliders down to 15%
Barren, Mountain and Wetland at 40%
All other sliders left untouched at standard 50 %

2 teams of 1 (start at war)
Ironman Mode: on
Seal Victory: off
Unifier Victory: off
Tactical Combat Mode – Against AI: Always Ask

Game speed: Normal (Fast Population after v1.32)
Starting Resource: Standard
Starting Skills: None
All Heroes Have Resurgence: Never
Defenders Strength: Strong
Cosmic Events: Medium
Maximum Number of Heroes: 2 (1 Leader + 2 heroes)
Maximum Heroes Level: 13
Map Exploration: On
City Founding: On
Random heroes match player race: Off
Empire Quests: On


Rules
1. The tournament is a championship with multiple sub-leagues and with promotion/relegation at the end of each season. Each sub-league has 4 players who play each other exactly once during a season. Each match played yields 1 point and a victory yields 2 additional points (so 3 points for a match won). At the end of the season, the best ranked of each sub-league are promoted to the sub-league above (if any) and the two worst ranked are relegated to the sub-league below (if any). If two teams are to a tie and the two teams have played each other this season, the winning team of this match is ranked above. For other ties, the ranking of the team is decided randomly.
One season lasts maximum 9 months. Players can start all their matches of the season in parallel if they so wish.
If neither of the players won 2 months after the start of a match, then the players should notify Judges who will take appropriate measures. Such measures can include asking one to surrender and take a screenshot of the score and provide it to the Judges. Based on the score, the Judges can decide that it is a draw and decide a replay the game with a shorter deadline (for instance 1 month), an extension or another solution (including awarding victory to one team), based on the circumstances of the match and the tournament.
In case of score victory, Judges may require the “new turn notification” logs to all players to determine if they delayed the game by repeatedly breaking rule 3. A malus (e.g. -10% or more) might then be applied to the score of the rule-breaking teams.

2. The settings described in the “In-Game settings” above are to be considered the default. If all the players involved in a game agree, they can arrange settings that differ from default settings in order to improve balance (e.g. if some exploit or imbalance has been discovered during the tournament but has not yet been dealt with the usual balance mod). Judges can oppose the settings change.
The player hosting the game is automatically and randomly chosen by the Battlefield website. If for any reason (e.g. no DLC), the designated host team cannot host the game, then the other team does. The host should start on slot #2. The host must take screenshots, preferably in windowed mode (showing the date and hour when the screenshots were taken), of the settings and post them publicly (so that his opponent and the Judges can access them) before the match starts.
No game restart due to bad starting conditions will be allowed except if one player is totally blocked (e.g. by Lava or Cavern Walls).

3. If one player temporarily cannot play his turns in 24 hours (e.g. because of holidays, or personal issues), and warn the host reasonably in advance, then the host should temporarily change the turn timer to accommodate him.
If a player skips his turn because of the timer, it is not allowed to revert the turn: the host should remove the turn timer and players should notify Judges and wait for their indications on how to continue.

4. In the event of a crash (from your PC or from AoW), the player who crashed must not re-enter the game after the crash, he must wait for feedback from the Judges allowing the game to continue: the player who crashed needs to take appropriate screenshots, pictures or videos (with a mobile phone for instance) and sends them to the Judges. The turn timer must be turned off by the host while waiting for judges’ feedback (see rule 8). Reloads should be counted by all players and reported to the judges. If no proof of a justified crash or in case of repeated reloading (during one match or during the tournament as a whole), the Judges have full power to disqualify the player.

5. Each season, a pool of 4 classes and 4 races available to all players during the season will be announced: during that season, players can pick their class/race only from this pool. On top of that, a player can never play a class or race that he/she played in all his/her previous matches of the season (so you can play Goblin only once in the whole season, and only if it belongs to the 4 races allowed this season, you can play Sorcerer only once, etc.).
BEFORE the match, you must let the Judges (but not your opponents) know what class and race you will play for your upcoming match by private message in the Battlefield website.

6. Banter and taunting is allowed, rudeness, swearing and racial (or other prejudicial) language is not. Repeatedly trolling will lead to forfeiting the match.

7. The decision of the Judges is to be respected. When the rules are imprecise or do not cover a case, Judges have flexibility to take the most appropriate decisions. Players arguing with Judges inappropriately get a total of 3 notifications before being disqualified from the tournament entirely. Judges may disqualify a player without notifications if their behaviour is deemed to be completely out of order (e.g. racism).

8. When awaiting Judges’ decision (as soon as his opponent requests arbitration), the host must remove the turn timer and may need to revert turns to come back to the situation existing prior to the issue being judged.

9. Players must play with the latest official version of the game (no beta). If the official version is updated during the tournament, players must update to it as soon as it is released both on GoG and Steam (and turn timer must be removed while waiting for the GoG release). Players must also play with the PBEM balance mod and Strong Defender Normal Reward mod. At the beginning of the tournament (season 1), v1.31 of the mod was used. All the games may be updated to a newer version if all the Judges deem it to be stable and necessary for game balance (they may require the non-binding opinion of the players through a poll).

10. If you feel one of the rules had been broken, but you carry on playing without reporting to the Judges, then you have no recourse.

11. Players should not voluntarily exploits (e.g. behaviours unlikely to have been expected by devs, unknown to the AoW3 PBEM community and/or which provide an unusual advantage such as unlimited healing, etc.), bug or weaknesses in the game. Players should report any bug or exploit to Judges or on the official forum when they discover them (and ask to Judges in case of doubt). Judges can decide to ban players using exploits from the tournament.


Arbitration process
The main Judge is Hiliadan.
Assistant Judges are: xx and yy

When arbitration is necessary:
1. If 1 or 2 of the Judges are involved in the game which needs arbitration, the 3rd Judge should name two temporary judges to replace them for this arbitration. If all 3 Judges are involved in the game, they need to all agree on a 4th person not involved in the game as a temporary Judge, which will then design 2 other temporary Judges; the 3 temporary Judges will conduct the arbitration;

2. The players should both send necessary files to Judges (screenshots, videos, their latest saved game which can be found here in Windows: C:\Users\[user name]\Documents\My Games\AoW3\Profiles\[profile name]\PBEM) by email or any other means;

3. The arbitration will be taken at the majority of the 3 Judges and the decision taken is definitive. If one player refuses it, he is disqualified.
 
Last edited:
I'm still unconvinced by keika and Rodmar's proposals.
Picking/banning for every match is usual practice in games and tournaments.
In some games you pick/ban maps from map pool, in other games you pick/ban hero or race.

In some games it is built-in, but for others people just chat and post voting results.
Sometimes people use simple tools/sites to simplify pick/ban process (just google few examples). There are absolutely no issues with that.

Current system is inconvenient for both sides - for judges and for players. Everyone needs to track previous picks for next picks and it is unnecassary and tedious.
I would say more, current system's bureaucracy is frustrating for players so that some players even abandon tournament because of that.
To be honest I don't remember anything like that in any other games/tournaments and when I faced it first time, it seemed very weird to me.

(and I am not even talking about playing weak class vs strong class just because you don't have any other options available)
 
@keika proposition is very simple, but need to be better defined to make it work. How many classes/race should be banned? Do we need to ban races? Should first player pick openly before second, so second player will pick with information against what he is playing, or both players are picking without knowledge from classes/races that are left in the pool. Maybe we could ban whole combination class + race and have more bans? And so on.
I don`t think that keeping track of race and class that you played is difficult, just write it down in any place and done. We are talking about 3 matches, if I understood correctly, it`s not rocket since to track that there is pool of 4 classes races and you must play 3 of them in those 3 matches. What is the point to track down it by judges, this tournament is purely for fun, if some one wont to cheat let him, what the point of controlling it?
@Rodmar18 proposition is very interesting, but some tool should be created to make it work.
 
What is the point to track down it by judges, this tournament is purely for fun, if some one wont to cheat let him, what the point of controlling it?
Unfortunately tournaments are not casual games and we know that some players do try to game the rules (and others just don't understand them properly and don't follow them, making it unfair to others), so it IS necessary that judges track everything to ensure rules are followed. So the system also needs to be easy to follow by judges.

For players, with the current system, tracking their (and others') past class/race choices is very easy: go to the Tournament's roster and open the past 1 or 2 games and you see their class/race. It's centralized and easy to find.
 
Here's a demonstrator that only implements a single player configuration, meaning that if it is used for several players in a row (as expected), their choices are not interfering with each other.
It is not entirely fail-safe, so if anything red is displayed, you just broke it.
 

Attachments

  • leaderDrawPBEM_demonstrator.7z
    18,3 KB · Views: 0