PDA

View Full Version : Faction war imbalance points to an underlying problem with the code



Nanowyre
02-24-2017, 09:52 PM
From what I see every day on these forums and ingame, there is an obvious problem with the faction war.

By all means, the faction war, as the factions are not tied to certain classes, should be a mostly balanced affair. That it strongly deviates from balance after such short time already is an indicator that there is an underlying problem.

Now, there are a lot of people here who argue "we are just better players". That is bollocks and you are idiots. My apologies.

You are basically a randomly selected third of the total playerbase. There is no way you are for some reason systematically better than the other two thirds as a whole by that large a margin.

This asinine "reasoning" has been heard in defence of similar situations over all manner of games for years. "The terran are just better players", "the horde are just better players", "the red team are just better players". It has never been true, ever. After time and balance, it universally turned out, that no, no randomly selected group will just magically be "better". In each and every case there were imbalances, often subtle, which ultimately caused the strong deviation from statistics - because all things being equal statistics win always. Sorry, Vikings.

Start looking for systematical problems with your code to adress this. Several pointers:

1) I have two Knight friends who every second game are unable to set assets because of an error. Apparently the assets are then set automatically, but who knows. It occurs for nobody else. It might be an accident that these two just happen to be Knights, or maybe there is a problem with asset allocation for the knights faction.

2) The way the territiries are set up may disadvantage certain factions. Set up a simple Monte Carlo sim, let it play 1000 faction wars with equal win probabilities for every faction, see who wins the wars. If it is not one third each your problem is with the territory setup. I would do it myself if I had more time.

3) Is there mechanism which accounts for numbers imbalance between the factions? If so, check that. Again, do a simple sim (after you checked the territories), but let one faction have twice the players of the other two. Keep the overall win probability exactly equal. See who wins. If the total population affects the win percentage, fix it. We are not here to play "there is more of us than of you". Or less, whichever way it turns out. Maybe it oversteers.

4) Check how your system handles load imbalances. Assume that the number distribution and the win rate is equal, but knights mostly play against samurai, samurai against vikings and vikings against knights. Do a sim. See if this breaks something in the statistical evaluations, maybe some weighting factors break if one faction only ever plays against one other. Then do it again in an asymmetrical case: let samurai fight vikings and knights fight vikings, but suppress matchups between samurai and knights. See if that breaks something. This scenario could happen if the factions are not equally distributed between time zones, which appears to be the case.

Do that, and with how the map looks now I guarantee you will find a problem.

Until then the state of the faction war is meaningless and your wins/losses mean nothing. Sorry, you are not magically better as a faction. You are winning like now, overwhelming odds are you are riding a bug to free points.

Get to it Ubi.

Valtaya
02-24-2017, 11:39 PM
If I understand it right, there is an autobalancer, however that only "seem" to account for total members of a faction.

The problem is, that is not going to work. They need a system that actualy accounts for "faction members playing".

You can see the autobalancer in action after each zone update... samurai start with ~20% more assets and vikings ~10% then knights... this is simply because there "seems" to be more knight faction members around, maybe even online... but when it comes to playing... it is not the case.

As you already said, the chance samurai and viking are SOOOOOOO much better players then knights is actualy ridiculous. Even IF they were SO much better, they would not be taking knight territory so fast (look at what happens in the night, CET). I tend to make a screenshot before I go to bed... 12-18 hours later (thats 2-3 updates), samurai got 2-3 more rows from knight territory and vikings usualy 1 line... how is that possible? When I play, I actualy see 50% knights or more... so they should have more assets, even if they are worse players (which is not the case, usualy knight players rank the highest in any match I am in). Well, by day more knights are playing... but have to fight the autobalancer first... by night we have more samurai playing, thats why they can not be stopped.

I think the system have to change from autobalancer from number of members per faction, to members who actualy played.

Lets say... each time a match ends the server notes who played, war assets are not important. By an zone update the server who handles the faction war knows (just example numbers and surely not realy realistic): 100.000 machtes played with a total of 600.000 players participated... of which: 300.000 knights, 180.000 vikings, 120.000 samurai. Lets further assume each player can deploy 1 war asset. Now, that would not be realy fair for vikings and samurai, understandable. Therefore we have the autobalancer... but that one acutaly works based (most likely) on total faction members and not players who actualy realy fought it out.

So, the autobalancer have to vanish and instead we get a distribution by "players who played". To have some lower numbers that are easier to understand. We have only 8 players playing For Honor
5 Knights
2 Vikings
1 Samurai
Lets further assume all players are of equal skill and by the end of match, their war assets are similar, not equal but similar. Even if your one Samura is the best player of the match, he will never be capable to deploy as many war assets like the knights do (except all knights just stand idly by whle the one samurai kills them all). As above said, we still assume each participating player can deploy 1 war asset after a match. After the match we have the following numbers:
5 Knight war assets
2 Viking war assets
1 Samurai war asset
Whatever you do... after a while Knights will conquer everything... so something have to be done. Now, how about a multiplier based on players playing.
5 is the highest numbers of a faction, so take that, we now divide it by the numbers of players from each faction.
Knights 5:5 = 1
Vikings 5:2 = 2.5
Samurai 5:1 = 5
That are our multipliers... we now take the war assets and multiply them.
Knights: 5*1 = 5
Vikings: 2*2.5 = 5
Samurai 1*5 = 5
Voila... and thats it, we accounted for the "uneven" numbers of players in each faction.

Now we know, you are not going to get 1 war asset... it is based on your performance... so it is more like 0.7 and 0.8 and 0.9 and maybe 1.1.... or realisticaly 1680 860 140... and so on. Samurai can deploy more assets per player, while Knights deploy more assets because of their numbers... and this way, this autobalancer can just go to, eh, hell.

Right now it is insane, 2 factions have actualy to fight the autobalancer when everyone is online... and even when there is noone online, just because of their members. The reason for the autobalancer why something is necessary is simple... Ubisoft can not force any customer to play a specific faction and they can not force anyone to play multiplayer. So maybe Knights have the highest numbers (membership), but just on the paper.

With the system as described, a faction will win, if they are good enough. Well maybe Samurai are realy that much better then Knights, but if that is true, the war assets will show. Well maybe Knights only play vs ai and everyone know that yields less war assets then pvp. If that is the case, so be it. But right now, we fight the autobalancer which is biased and surely not in favor of the Knights. Right now it sickens me a lot... Samurai seem to play during CET night... assumption? Samurai are americans or asians, while most kngihts are european. Right now, I know a lot of Knights who just nolonger give a "s***" about the faction war and do not bother vasting time deploying war assets... why for? If you see how after a zone update Samurai have like most new zones at 30%:70% in favor of them... why even bother? And then we also have the problem of not seeing what actualy happens and what is our impact. The other day I had a good run and deployed 20.000 war assets that evening... zone percenage have not shifted 1%. If you actualy do not know where your war assets are usefull and where just a waste of time, what is the point?

TheEpicRage
02-24-2017, 11:55 PM
A lot of people seem to like to blame the system and it very well may be the case but personally I think it comes down to one main thing at the moment.

Luck.
Now you may be thinking how is luck involved? Well I don't mean anything like lucky wins or such but luck in how many people have either stumbled on or have had the luck to be told they can deploy assets in areas outside the playlist they are playing in.

You would be surprised how many people don't know this.

I've noticed a trend that wherever the dominion playlist sits tends to be the are with the most territories swapping hands.

TheEpicRage
02-25-2017, 12:23 AM
@Valtaya
I'm pretty sure the autobalancer changes asset gain by current active players at the end of each game, I mainly do duel and brawl and do consistently well and see pretty substantial fluctuations in asset between games.

I assume this is why occasionally the game has trouble with post game rewards and waits to be able to check in with the servers before giving out assets.

xXl Plan B lXx
02-25-2017, 12:32 AM
If you have some guy behind a curtain making decisions that you can't see, it's much easier to point your finger at the hidden things happening. If there were a much clearer representation of what the faction war is doing, how it works and what makes it change then I think people would be a lot less skeptical of what is happening. I personally know that I've put about 175K assets into the faction war. That's probably pretty low, I don't know as I don't have a ton of friends who grind this game. I feel like that numbers just monopoly money when I log in and see an image that shows that the knights have 9 territories.

http://i.imgur.com/VTS4NSa.jpg

For example this is the current map. How do the knights have 31.2 % of the assets yet have almost 50% less territories than the Samurai?


A lot of people seem to like to blame the system and it very well may be the case but personally I think it comes down to one main thing at the moment.

Luck.
Now you may be thinking how is luck involved? Well I don't mean anything like lucky wins or such but luck in how many people have either stumbled on or have had the luck to be told they can deploy assets in areas outside the playlist they are playing in.

You would be surprised how many people don't know this.

I've noticed a trend that wherever the dominion playlist sits tends to be the are with the most territories swapping hands.

Aeecto
02-25-2017, 12:44 AM
A lot of people seem to like to blame the system and it very well may be the case but personally I think it comes down to one main thing at the moment.

Luck.
Now you may be thinking how is luck involved? Well I don't mean anything like lucky wins or such but luck in how many people have either stumbled on or have had the luck to be told they can deploy assets in areas outside the playlist they are playing in.

You would be surprised how many people don't know this.

I've noticed a trend that wherever the dominion playlist sits tends to be the are with the most territories swapping hands.

Nothing to do with luck.... the autobalancer favors samus and viks and seems buggy or is realy badly coded for knights.
Viks and samus had both a nice boost once they were behind (initial territory control each map-update already at 52-53% and knights starting at 47-48%... and a huge boost for war assets gained), but the knights havenīt got it yet even getting steamrolled over night (CET).