134 replies to this topic

[ScarecrowES]

Alright, so I've done a little thinking about the root cause of the Faction Warfare Tug-of-War imbalance problem. I've been following the progress of PGI to bring some balance between the performance of both factions in FW - now through two hot fixes, drop deck tonnage changes, and so on.

The data PGI has been collecting has seemed to be well out of step with the state of every capture session. PGI says that the Clans are not performing, match-to-match, significantly better than the IS, and in many cases worse... and yet, the tug-of-war stays pegged at max to the Clan side in nearly every moment.

After doing some real analytical thinking, I had to facepalm myself when I realized that it literally has nothing to do with Clan vs IS balance, or faction skill distribution, population distribution, drop deck tonnage distribution, etc.

The reason the tug-of-war pegs out to one side of the conflict so readily is because of a inherent design flaw as to how match results influence the tug-of-war.

PGI chose to have every match move the tug-of-war peg by a fixed amount. This, more than anything else, is the cause of all FW 4.1's ills. This means that a faction needs only to maintain a specific win amount differential over the other faction to keep the tug-of-war pegged to max, rather than requiring a significantly higher win percentage to do so. Explained in detail:

Each win in FW moves the peg in the tug-of-war 3.33%, regardless of how many games have been played. This means one faction only needs to maintain a 30-game win differential to keep the peg maxxed for that faction. To move from a completely neutral position to a full-max position, the peg only has to moved 100% (100 points). If each match is worth 3.33%, then it only takes 30 more wins than losses for one of the factions to reach max.

Reaching a 30-win differential would be difficult if we're talking about a series of 100 matches. That requires the winning side win 65 of 100 matches - almost 2/3. But over a series of 1000 matchs? The winning side would only need to win 515 of 1000 matches to reach a 30-game differential. That's not even 52% of matches won! The more games that are played, the lower the win percentage needs to be for the victor to reach the required win amount differential.

This means that the tug-of-war will ALWAYS be a stomp for one faction or the other - even if every factor of balance was perfectly tuned. As long as each win is worth a fixed amount, this will always be the case.

It also means that the only match performances that really matter are those that occur in the final minutes of a session, where the defending team has a chance to produce the very SMALL win percentage necessary to deny the attacking side a cap.

It has absolutely nothing to do with whose tech is better, or where the top-skilled teams are sitting. It has everything to do with the basic inherent design of the tug-of-war. If we hope to have any chance at Clan vs IS balance in FW, we MUST change how the tug-of-war works.

[SuomiWarder]

That is an interesting take on things, if the internal math of their system functions as you surmise. They would never admit to making such a blunder but maybe if it is true, they will make some quiet changes.

[ScarecrowES]

SuomiWarder, on 20 December 2016 - 05:30 PM, said:

That is an interesting take on things, if the internal math of their system functions as you surmise. They would never admit to making such a blunder but maybe if it is true, they will make some quiet changes.

It's not a matter of surmising. PGI has said as much. Copied from the Dec 13th Patch Notes:

"War Log tab

The new War Log tab provides a complete history of all changes in the Tug Of War progress for your selected Conflict. From here you can track all of the Tug of War changes throughout the course of this Conflict, such as which side won a match and when, and by what percentage that win altered the Tug of War. By default, all matches will affect the Tug of War by 3.33%."




As you can see, every win moves the peg 3.33%. As long as one faction has won 30 more games than the other, they will max the peg. Hard thing to do with 100 games... it means winning 65% of the time. Super easy thing to do over 1000 games... you only have to win 51.5% of the time.

It also means that the margin of victory is extremely small. One faction can maintain that 30-game differential over the 8-hour period, and all that has to happen for the defending side to deny them a victory is for the defenders to win a few more games than they lose in the final few minutes of a session.

For this system to work, balance has to be sooooo precise... and I mean, both Clan and IS tech balance, faction population, skill distribution, so on and so forth... so precise that neither faction can maintain a 30-game win differential over however many games played per 8-hour session. You're talking about being unable to fall more than a few tenths (maybe hundredths) of a percent of 50/50 in order for the tug-of-war to not be a complete stomp in every session.

Edited by ScarecrowES, 20 December 2016 - 05:43 PM.

[Andi Nagasia]

interesting, but that means the team that has the lead has to keep said lead,
if they start losing games it will start going back the other way as due to over flow,
(over flow meaning that if your Bar is Filled you winning wont increase your bar)

i suppose it could be changed to matches matter less the farther down the bar you go,
(if your sides Winning(your Bar is more Full) / (if your sides losing(your Bar is less Full)
-
each FP Skirmish match Victory nets +5% of the bar(if your sides winning),
each FP Skirmish match Victory nets +5% of the bar(if your sides losing),
-
each FP Domination match Victory nets +4% of the bar(if your sides winning),
each FP Domination match Victory nets +6% of the bar(if your sides losing),
-
each FP Conquest match Victory nets +4% of the bar(if your sides winning),
each FP Conquest match Victory nets +6% of the bar(if your sides losing),
-
each FP Assault match Victory nets +3% of the bar(if your sides winning),
each FP Assault match Victory nets +7% of the bar(if your sides losing),
-
each FP Invasion match Victory nets +3% of the bar(if your sides winning),
each FP Invasion match Victory nets +7% of the bar(if your sides losing),

this may fix that problem but it also may over complicate things,

[MischiefSC]

A 60 match variance would still flip either way, just on "luck" if balance was even moderately close. The critical issue is skill and population distribution. Because skill distribution in FW tends to build (the more competitive someone is the more likely they will go where they see an advantage and seek out the company of other skilled players) the issue grows over time.

If there's a better than minimal tech difference you will create an inevitable unit/experience drain to that direction.

The win/loss is nowhere close to 51.5%. It's about 80/20. While I get the concern on the tug of war mechanism at the moment it's not really relevant.

[Y E O N N E]

Interesting, though I'm not so sure the hubbub is really about what the progress bar says as much as it's about the match-to-match trends and how the games themselves are playing out.

[ScarecrowES]

Andi Nagasia, on 20 December 2016 - 06:03 PM, said:

interesting, but that means the team that has the lead has to keep said lead,
if they start losing games it will start going back the other way as due to over flow,
(over flow meaning that if your Bar is Filled you winning wont increase your bar)

Yes... you're essentially right, which is why the defending team can come in and sweep away 8 hours of success in a mere few matches at the end of a session. However, maintaining the paltry win differential required over that 8 hour session is ridiculously easy. As I showed, if only 1000 are played in 8 hours, the winning team only needs to win 52% of the time to keep the bar maxxed in their favor.

So long as we give each win a fixed percentage toward the bar, it will always peg at max. The only difference is how quickly we get to that max at the start of a session, and how easily the defenders can rally.

No... what needs to happen is that during all of the middle sections, the amount the bar moves needs to be based on a percentage score, rather than a fixed amount.

For instance... at the end of the match, each player on the winning side gets a point for their faction, and an additional point for good performance. The losing team can also get a point for good performance. So each winner can have between 1 and 2 points added to their faction total, and the losing side can get 0 or 1. You compare each faction's total points against each other, and THAT's where you put your peg.

Maybe once you get to end-game territory and are in the Invasion block, you can do pure win/loss as we get now to give the defenders a chance to rally.

But so long as each win is worth a fixed amount of the total, you'll always tend to be pegged in Invasion, no matter what the game's balance is like.

[ScarecrowES]

MischiefSC, on 20 December 2016 - 06:10 PM, said:

A 60 match variance would still flip either way, just on "luck" if balance was even moderately close. The critical issue is skill and population distribution. Because skill distribution in FW tends to build (the more competitive someone is the more likely they will go where they see an advantage and seek out the company of other skilled players) the issue grows over time.

If there's a better than minimal tech difference you will create an inevitable unit/experience drain to that direction.

The win/loss is nowhere close to 51.5%. It's about 80/20. While I get the concern on the tug of war mechanism at the moment it's not really relevant.

The point of getting the tug-of-war right is that it both determines which modes you play, and how easy it is for one side to dominate the other.

The current system basically guarantees you're going to play invasion most of the time, as the most likely state of the tug-of-war will be for one faction to have the bar maxxed. The other issue is, that makes one side overwhelmingly guaranteed to win the session. The only way they won't be guaranteed a win is if the defenders rally in the final few posted matches.

Otherwise, the system will always place extremely lopsided results, regardless of what balance is like. It's a foregone conclusion... whichever side can maintaint even a tiny win differential is almost guaranteed to win the session.

And so far, PGI says match results are a LOT closer than you'd expect... more like 60/40 than 80/20. 60/40 is MORE than enough if each win moves the bar 3.33%. It wouldn't be enough to guarantee a victory under other methods of moving the peg, though.

Edited by ScarecrowES, 20 December 2016 - 06:22 PM.

[JaxRiot]

Edit

On second thought, I think I will stay out of this one.

Interesting thread though

Edited by JaxRiot, 20 December 2016 - 07:09 PM.

[ScarecrowES]

JaxRiot, on 20 December 2016 - 06:53 PM, said:

This thread is actually kind of interesting.

So this means that if one side can keep a 4 game winning lead per hour, then that faction will keep stay pegged.

But do we know how many games are actually being played per hour on average?

I ask because the the Tug Of War gets maxed out at 100, so we wouldnt be able to tell if one side was winning 52% of the matches or 80%. We would just know that one side is keeping a 4 match lead.

For the sake of clarity, it would take a 30-game win differential at any point over the session to peg it at Max. The more matches in a session that are played, the easier it is to reach that amount. Obviously wins past that 30-game differential don't count, as you can't go over 100%. Once a 30-win differential has been achieved, the attacking side merely needs to maintain a positive win percentage to keep the peg maxed.

If one faction has even a micro-advantage over the other, it will be able to post a win percentage well above what is required to achieve a 30-win differential and maintain it over 8 hours. Even a tiny advantage will translate to a higher win percentage, and as I've shown, at 1000 games you need less than a 2% win advantage to dominate.

[JaxRiot]

ScarecrowES, on 20 December 2016 - 07:11 PM, said:

For the sake of clarity, it would take a 30-game win differential at any point over the session to peg it at Max. The more matches in a session that are played, the easier it is to reach that amount. Obviously wins past that 30-game differential don't count, as you can't go over 100%. Once a 30-win differential has been achieved, the attacking side merely needs to maintain a positive win percentage to keep the peg maxed.

If one faction has even a micro-advantage over the other, it will be able to post a win percentage well above what is required to achieve a 30-win differential and maintain it over 8 hours. Even a tiny advantage will translate to a higher win percentage, and as I've shown, at 1000 games you need less than a 2% win advantage to dominate.

Ya, I tried to edit out my post and bow out of this one because after I read what I posted, I noticed that my finger brains didnt translate what my head brains were trying to say very well.

But I guess now Im in...

So what I am trying to say is that, unless we know the actual numbers of wins, then all we really know is that one side was able to gain at least 30 more wins than the other.

I mean, sure it makes mathematical sense that if 1000 games are played, then all the other side needs is less that a 2% win advantage to dominate, but we dont know for sure if the games are actually that close because we dont know the actual numbers.

Clans could actually be winning 30 out of 40 games for all we know, or more.

Edit- As interesting as the OPs take on the Tug Of war system is, Im with Mischief on this one.

Edited by JaxRiot, 20 December 2016 - 08:02 PM.

[BLOOD WOLF]

This is nonesense. No correlation.

Edited by BLOOD WOLF, 20 December 2016 - 07:38 PM.

[Artgathan]

That's certainly an interesting analysis of how the design of Tug-Of-War system lends itself towards stomps.

It's certainly not an explanation of why Clans tend to be the ones doing the stomping.

[Mystere]

In Hades' name, let's just have a real planetary campaign system already and none of these highly simplistic planetary invasion resolution methods.

[BearFlag]

Victory ultimately deriving from some micro-advantage (preferably a player's skills) is precisely what you want in game design. What you don't want (usually) is advantage based on start condition, luck or other imbalance. There are many types of games, of course, and some do use luck to mimic system complexity beyond the reach of a simulation's detail (or just for fun in a child's game). Chess has "first move advantage" which, even at the highest levels of play, results in white victory ~53% of the time. For this reason, chess is played in tournaments. A 60/40 split is just unacceptable in any game. The Holy Grail is 50/50 so that human agency determines the end game.

The 3.33% blocs are just chunks on a primitive map. It would be the same on a planet map with, say, provinces. The difference is that in other games (wargames, or Risk for that matter) victory conditions are met and the game ends. In Risk there is no Atlantean continent from which to continue the war indefinitely. It is rather more the lack of a proper game (phase) end than a math problem which is design flaw. But early phase end would speed the strategic game up even more. As it is, important things like a planet battle should exceed the circadian period by 2 or 3 times. But you need real balance.

I'm sure there are a number of ways to handle the pegged needle. But would they be any better in the absence of balance and of planet battles? The tug-o-war abstraction is something of a bass **** any way.

EDIT: For the sake of definition, by "balance" and "game" here I'm referring to a match. The larger picture should ~unfold~ instead of waking up to eight more dots changing hands.

Edited by BearFlag, 20 December 2016 - 09:36 PM.

[BLOOD WOLF]

This is really a non-issue that people are going to blow out of proportion.

[Y E O N N E]

Again, I do not think it is the overall strategic game causing the complaints. It is that, within a single match, one team has a starting advantage more often than not and it is resulting in lopsided matches that are not fun to play. If the individual matches weren't routs as often as they are, there would be far less complaints even if one team was still winning at the higher level meta-game, which is entirely irrelevant.

Edited by Yeonne Greene, 20 December 2016 - 09:34 PM.

[HellJumper]

BLOOD WOLF, on 20 December 2016 - 09:30 PM, said:

This is really a non-issue that people are going to blow out of proportion.

then why not counter the op in a similar manner he has done? (providing numbers etc to say this is a non issue or has no correlation or is a non sense) that can help you a lot in my opinion

[Novakaine]

We need a real campaign.

[ScarecrowES]

Artgathan, on 20 December 2016 - 07:46 PM, said:

That's certainly an interesting analysis of how the design of Tug-Of-War system lends itself towards stomps.

It's certainly not an explanation of why Clans tend to be the ones doing the stomping.

Of course. At no point am I saying that we have perfect balance between the factions, populations, skill levels, etc etc otherwise irrespective of the system in place. I don't intend people to read this and say that PGI should just give up on trying to find the perfect balance possible for the two factions.

What I'm saying is, under the system that we're operating under, even the smallest imbalance will inevitably lead to one faction or the other completely dominating the tug-of-war. The needle will always be pegged to the maximum in one direction or the other by virtue of the chosen design of the system.

And ultimately, that's absolutely bad for the game mode. The entire functional premise of the mode is two-fold... 1) that the mode represents a real tug-of-war with two factions that - all other things being equal - have an equal chance at winning, with the winner being the faction that simply outperformed the other... and 2) that players progress through a variety of modes along the path from neutral conditions to win conditions that will produce a variety of play scenarios and a sense of actual progress.

Unfortunately, with the current system, the result is largely predetermined for the faction which currently has any advantage over the other - no matter how minor. In this case the Clans have that advantage. The system is fundamentally set up to expound on that advantage. And by nature of the tendency for the system to peg the needle at maximum for the advantaged faction, we're completely unable to play through any of the other modes or engage in a real tug of war.

Ultimately, the manner in which wins affect the tug-of-war has to change. You can't even BEGIN to look at faction balance so long as there is such a glaring flaw in the base system.