There will be light/medium/heavy versions of every weapon and every robot at some point. Just give it time.... LOL

As with other pilots, I have my share of issues with the game. I'll share those in other comments, if you'll let me explain.

It's hard to take MAIL.RU (Pixonic) serious when only platitudes are given. This is evident in every aspect of how MAIL.RU operates, beginning with policies on this website. In particular, Section 6. USER CONTENT in the terms of service, which in part reads:

"Pixonic reserves the right in its sole discretion to review, monitor, prohibit, edit, delete, disable access to or otherwise make unavailable any User Content (including without limitation your User Content) without notice for any reason or for no reason at any time. "

This clause tells me that there is no serious effort to address issues. In fact, as someone who has had no less than 10 accounts locked because of my comments, none of which were inflammatory, racial, attacking other's commenting or in any way violated any of the enumerated reasons for locking accounts, I can say with a high level of confidence that anyone who makes commentary that is too close to being the truth on how Pixonic operates gets their accounts locked so they can not continue raising visibility to the issue.

In one of my prior comments where I discussed matchmaking in length and why the old system was better I jokingly said at the end of the commend, "How will Pixonic respond? DELETE-DELETE-DELETE...and this is how we deal with dissent in the USSR..off to the Siberian gulags."

Guess you guys don't have a sense of humor, cause it was in fact deleted.

I find it curious that after many months of deleting content and suspending/locking accounts that you're now responding to the comments by the community. I can only surmise that this is because you have been directed to do so because the "vibe" is so negative now that management has changed their policy from "no public interaction" to "positive assurance interaction". Meaning you've been instructed to repeat what has been posted on facebook and the blog, but improve the P.R. image by personally responding to the concerns.

Let's see if you'll keep THIS account open long enough for me to (again, sigh) raise a couple serious issues the majority of the community has with the game and why many players and former players feel that Pixonic's failure to take action demonstrates no real interest in addressing the problems.

Good Morning,

I will get into what many see as the biggest single issue of the game toward the end of this comment. Before I do that, I have a question that I have asked on numerous occasions and never gotten an answer. Not just me, for if you look at many of the comments on facebook, it's not even a question anymore. Because of Pixonic's refusal to directly answer the question it's taken as fact by most of the community due to the company's silence on the matter and actions which indicate it is a correct assertion.

The question is this: Is War Robots now being run under a PAY2WIN model?

Personally, I believe it is and I'll give my evidence to support that position.

I will define PAY2WIN so we're using the same terminology. There's quite an extensive wiki entry but I'd like to focus narrowly on the core of the matter as it applys to WR. PAY2WIN allows a player to buy new equipment that provides such an advantage over any player who has the best available old equipment that no amount of skill or strategy can overcome that advantage.

To dissect that, one really needs to look at the data and do some statistical analysis. Since I don't have that, and for the purpose of argument, let's do a simulation.

Out of all the players on WR, say the best player skill wise hits their target 80% of the time. Let's say the worst player skill wise hits their target 50% of the time. The difference in the damage they inflict is 30%. Let's say they both happen to have identical equipment (best MK2 available in terms of dealing damage) and over the course of a battle the total amount of damage they can inflict is 1,000,000

Best player - 800,000 damage during the game
Worst player - 500,000 damage during the game

Now lets say Pixonic makes a new weapon available, like, say, the Shocktrain. The only way it is immediately available is to pay cash for it. Over time it can be won through components and buying components with gold, etc, but that will take months.

Let's say the shocktrain can inflict 100% more damage during the course of a battle and the Worst player is tired of getting trounced because he is the worst player, so he gears up with shocktrains. This completely changes the outcome.

Best player - 800,000 damage
Worst player - 2,000,000 (amount of damage ST can do during the game) x 50% = 1,000,000

This is PAY2WIN. By having a weapon immediately that is only available through purchase, even the worst player can defeat the best player. If that's not bad enough, the impact of the ST in the hands of the Best player is absolute in terms of impact on the game (good for those who have it, bad for everyone else).

Let's look at the evidence of what Pixonic has done that supports the position that this game is now PAY2WIN.

1. Releasing overpowered weapons and robots that give an extreme advantage to players to have the gear as I described. Referencing Bradley Snow's comment (which you already responded to). There is a test server. Using that test server, and the example I've given, I will reiterate what Bradley has said. How could the developer's not know?

2. Minor tweaks and re-tweaks that maintains the weapons overall superiority, but closes the gap with existing weapons marginally (Nerfing).

3. Every couple game release again introduces overpowered weapons and robots.

4. Despite all the platitudes on addressing the problem given in the news and posts (here and facebook) and the functional changes in each release, the net result is minimal. The newest released weapons and robots are magnitudes better than the older gear and overcomes any deficiency in skill a player may have.

5. THE IMPLEMENTATION OF A MATCHMAKING SYSTEM WHICH ALLOWS MANIPULATION OF THE MATCHMAKING OUTCOMES, RESULTING IN GROSS MISMATCHES.

Number 5 is in caps because as this is the core issue with the game. If this core issue were addressed many other issues would completely go away or those remaining issues would be the equivalent of an irritating mosquito bite instead of the compound fracture of the leg that they are today.
--------------------------------------------------------------------------------
To reiterate: Is the net outcome that this game now PAY2WIN? Based upon my observations of the history and direction of the game, I believe that to be true. So the question is why?

I can only come up with three possible explanations:

1. The person calling the shots on game features and functionality is inept. He/she is making poor choices leading to undesired outcomes.
2. The data and processes used to make the decisions on features and functionality are lacking and broken.
3. The policy on the game is in fact PAY2WIN.

I've probably covered enough for this comment. I won't get into the matchmaking issue unless I can get a definitive yes/no answer on whether the business model is PAY2WIN. The reason being that I don't want to waste your time or my time. If the intention is to run the game as PAY2WIN then any discussion around matchmaking is frivolous, since fixing it would be counter-productive in supporting a PAY2WIN strategy.

You're conflating (as does Pixonic's Terms of Service) two separate topics; user game play and user comments/posts (both here and on facebook) . I can assure you that this is not standard for any IP. In fact it is juxtaposed to the policy of one of the biggest IPs and I will provide the proof that backs my point. My point being this: Pixonic's policy is vague, capricious and is used to stifle any criticism or conversation critical of the game or Pixonic's perceived modus operandi.

The Terms of Service I will reference are from EA. You can reference them here (http://bit.ly/2HwYDUD).

There are a few things I'd like to point out that differentiates their policy from your.

1 They enumerate exactly what is not acceptable in an enumerated list. One of the items in that list highlights the difference between your policy and theirs:

"Publish, post, upload or distribute any content, such as a topic, name, screen name, avatar, persona, or other material or information, that EA (acting reasonably and objectively) determines is inappropriate, abusive, hateful, harassing, profane, defamatory, threatening, hateful, obscene, sexually explicit, infringing, privacy-invasive, vulgar, offensive, indecent or unlawful."

The key segment there is "acting reasonably and objectively".

2. They have a specific procedure for suspension of privilege of posting/commenting and locking/deletion of an account. In effect, they give the person commenting the benefit of the doubt and will issue a warning first, unless the content was so egregious that it merits the content being removed and the account being locked/terminated immediately.


I won't go back and forth on this. I have presented one example of what I think is solid evidence backing my position. I'm sure I can find others. Pixonic can choose to ignore it or take specific action based upon the specific feedback and information provided to improve the customer experience (in this case specific to content management for posting on this website, censorship and account suspension).

Moving on, my next comment will be specific to the core problem of the game and Pixonic's failure to act which has caused me to myself and a large number of my fellow pilots to come to the conclusion that no real effort is being made to act in good faith and correct the issues.

What level robots do you have?

Understood on the staffing given the amount of time you've been with the company. Hope you know someone to pass that recommendation over to along with why I made the recommendation.

You may want to pass along that I really didn't have any bias one way or the other on the P2W question as I just wanted to know the position of the company. So for him to say he is sure I didn't want to hear that, my response is he was sure incorrect.

Moving forward. Matchmaking. Where to begin

I think I need to understand your position, because to effect change I (or a collective we for all of us the who think the current system is garbage), need to have an advocate inside Pixonic to support us.

So, a general question. Do you agree or disagree that a matchmaking system where the outcome can be manipulated is foundationally better or worse than one which can not be? I ask that you don't interject any flaws in the current system or the old system to bias your answer.

OH, and I've used the gala had and even Gareth to go against shockers, unfortunately teammates see you going one on one and try to join in. It's all over for the Gareth in an instant for the Gareth and in a handful of shots for the gala had because you get hit from behind on the ricochet. BTW, lightning fans out on a strike, it doesn't go 180 degrees after it hits a target. Kinda counter to physics and natural phenomonon don't ya think?

Hmmmm. Interesting reply, for several reasons.

If you're going to the dev team to get an answer as to whether the game is now positioned as PAY2WIN, I have my concerns about how policy is being decided in the company (unless one of the developers happens to be a majority owner or has a dual role and is also CEO or CTO). I would have expected the answer would come from someone higher up in the organization chart hierarchy.

Good to know that you're also a player of the game. Helps out significantly in knowing that you're active in the game and not someone who was brought in just because they met the skillset requirements in the job listing.

I don't want to suggest that players should expect that they will never need to grind it out now and then. Nor to say that we shouldn't drop some cash to instantly upgrade or buy the good robots or weapons if one has special appeal. Obviously without the purchases Pixonic won't survive. That being said, there is also no reason for Pixonic to release weapons and robots which are so over the top that it ruins the gaming experience for every person who doesn't have them.

Respecting your comment and the situation you presented. Agreed, strategy would come into play. However, what I described is a situation where there is so much disparity in the amount of damage that strategy can not over come it.
Let me re-emphasize, if the difference in firepower and damage is great enough it will overcome any strategy and skill difference. The outcome is assured. It will come down to math without exception. Let me expand on your example a bit more to better illustrate.

You mention the Fury with Zeus as the counter to a dash with shocktrain. Funny you should mention that, as I have a Fury and have gone against shocktrains with an ancil and 2 zeus' and with 3 zeus'. Both the robot and weapons are and have been at level 6 for quite some time (saving for something better). So what is the outcome of situations where I come up against a dasher with 3 shocktrains?

Well, for one, you can almost completely negate that 100m distance advantage for the zeus. First reason is because of the differnce in speed between a dash and Fury, the dash can quickly close the distance to get within it's range or increase the distance to get out of my Fury's. So the Fury usually gets 1 and on occasion 2 shots where the dash can't fire back. Also, there are only 2 maps that provide the long, clear shot distances where the dasher can't easily find a spot to hide behind (and with TDM there are practically zero!). This provides an advantage to the dash. Then of course, the shock can hit you if there is another target within 150m of you, so you very seldom have that 100m advantage because you need to position yourself where someone cant quickly sneak up on you since the Fury is so slow, yet you still have the distance to other targets to inflict damage where they can't do the same. That means you'll almost always have teammates less than 150m in front of you.

Although the zeus has a damage advantage over the shock and can fire faster, firing faster actually hurts the Zeus-Fury combination and results in reducing the firepower advantage if there is cover for the dash (and given their speed and map layouts, there usually is). The zeus fires every 6 seconds. The shock, to a full charge, takes 8 seconds. The effective firepower of the zeus is reduced by (100% - 6/8% = 100% - .75% = 25) 25% (again, it's the math) because the dash pilot would/should know that it is to their advantage to fire off full charges and then cover. In the next release, by making this time to full charge longer, in this instance it's actually working to the shocks advantage. Did the devs think about that? Probably not.

Actual game experience summation. I can take out a dash with shocks that are up to level 9. Up to level 7 dash with shocks I can get most the time, but level 8 becomes much harder and level 9 is very difficult. Once the % difference between skill and strategy is overcome by % of firepower difference, firepower wins, every time. The point of my previous comment was at a certain point where that difference is so great, it comes down to math, every time.

This is why matchmaking is so critical.

An MK2 dash with MK2 weapons (shocktrain, orkan, taran) will seal club even the best suited level 6 Fury configuration. Even the worst player is going to figure out that to take out the Fury they need to close the gap to get a shot. Problem is many of these dash/shock combos are by tankers, often in clans who range in skill from very good to average (seldom poor gaming skills). Because of the way shocktrain works, there are multiple ways to take out an opponent. Matchmaking, ugh! In it's present form it exasperates and magnifies all the issues. Let me emphasize, ALL OF THEM.

If Pixonic doesn't have a mathematician and statistician on staff, that is a big problem. It would also explain why stuff comes off the test server half-baked. The proper simulations and analytics haven't been done to provide accurate estimates on the outcomes in your production environment.

Anxious to hear back from you on whether or not Pixonic's strategy for the game is now PAY2WIN.

Have a good afternoon. And sorry you never got that lance. Now that you're with Pixonic maybe they'll give you one as a bonus.

" I can explicitly assure you that you have a lot of people in the general players' corner than you know who constantly provide feedback... *
am more than a bit confused here. To say one is in a corner came from boxing. With respect

Fixing Matchmaking - 2: Weapons, Damage, Damage Spread & Nerfing/Buffing

Nerfing/Buffing
The net effect of nerfing/buffing weapons is really more of an issue because of the way the current system works. Using a system based upon potential damage of a weapon during a battle makes nerfing/buffing immaterial. In fact, matchmaking using the amount of damage calculation would completely eliminate the need to nerf/buff.

How?

You could include in the table that has the total damage calculation a weighting variable. By default, all weightings is 1 (100%), but if you need to nerf or buff a weapon you change the weighting. So if you want to nerf a weapon, change the matchmaking weighting from 1 to 1.1 (this will increase the MM damage score so it is matched against hangars with stronger weapons. If you want to buff a weapon, change the weight to less than 1 so you match up against hangars with weaker MM damage scores.

However, there are things concerning damage and damage spread that will impact the calculation and how fast matchmaking can work it's magic.

My opinion is that regardless of the MM system, weapons and robots should not be nerfed unless nerfed equally. Ever. It creates too much community dischord and negative feedback and P.R. for Pixonic.

If you absolutely muct nerf a weapon, buff all other weapons. Instead of nerfing weapon "A" 10%, buff weapons B-Z 10%. Yes, you're doing the same as a nerf, but it comes down to perception.

Damage Spread

What I mean about damage spread is that if we were to look at the amount of damage each weapon has at a specific level, that the difference should be such that the spread isn't so great that it makes matchmaking take an unacceptable amount of time.

Point being that if I have a pool of 30 weapons to choose from and the amount of damage from the strongest weapon down to the weakest is 500% at level x, then you are really going to have a wide spread of damage values for calculating matchmaking, causing extreme segmentation of the active players, which will require MM to keep increasing range in the amount of damage to find matching players.

Keeping it tight not only within weapon class (heavy, medium, light) but across weapons should be by design.


Damage

The last piece of the MM formula, and probably most critical is the amount of damage each weapon can cause and how that is calculated.

Static Damage Weapons: For the weapons that the damage is static (magnum, punisher, spiral, zeus) the calculation is easy. Total the amount of damage for each round/charge for each weapon.

Distance Damage Weapons: For shotgun type weapons such as Thunder, Storm, even scourge. For these weapoins the calculation is a bit more complex but still not that difficult. But the most equitable method is to take the amount of damage a weapon can do at mid-range. So if a weapon does 450 at point blank and 50 at it's maximum range, use a damage calculation of 250 = (450+50)/2. This rewards the players with good WR skills who can get close to their target before getting destroyed. Statistical analysis of actual gameplay might yield that the weighting or "average" for the damage score is being considered for change. But I would try to stay away from doing that. Doing so you risk punishing players who are aggressive and skilled at closing on the enemy.

Variable Damage Weapons: This is for shocktrain and other similar weapons where it's damage is not dependent strictly on the player's skill or strategy and where damage is determined by play situations and scenarios. I think the best way to calculate damage is purely through statistics. Taking a look at the damage the weapon causes, most damage, minimal damage (minus tanker) and look at the statistics for the mean and within .5 standard deviation of the mean. Even doing that, you're likely to have a very broad range in the amount of damage at each level.

As I said before, Pixonic has really created a mess with shocktrain. The dash/shocktrain combination of this matchmaking systems is the equivalent of the gephard-magnum combination of the old system. Ironic, isn't it?


I have a few other thoughts that aren't top of mind at the moment. When they come to me I'll send more feedback.

On the P2W question, it is difficult to reconcile that the manner in which Pixonic has been releasing and making available new robots and weapons which certainly establishes a P2W gaming environment with what FM9CUY said when I asked the question about the policy and the answer being that the game is not P2W.

Speaking professionally, I can convey that I see this happen regularly, where the outcomes don't align with policies or objectives. However, where that happens is usually in large, multi-national companies that have been around for decades with thousands of employees and a myriad of corporate culture and political baggage which causes the misalignment.

I don't believe that's the case here though because FM9CUY has said they have all this data they analyze and the test servers, etc. So it is possible they just have a myriad of broken processes that need fixing.

Still, that is not an excuse for taking so long to address the issues if they are sincerely trying to fix them. There is no sense of urgency and that is wearing on the user community. My opinion. If Pixonic spent 1/4 of the effort on fixing matchmaking as they did on new robots and weapons this would have been fixed inside of three months.

Pixonic should have a sense of urgency about this because the bad gaming experience has a cumulative effect. Frustration leads to disgust, to sarcasm, to cynicism, to negativity and ultimately disconnecting. Based upon my observation, I think the majority of the community is in the last two or three categories. That's not good for Pixonic as it has a accelerating effect toward the last category.

Unfortunately, the people who aren't leaving the game are the same people who are making the gaming experience terrible. If I were to guess, it would be that most of them are also the P2W players. They have the best weapons and will tank down the leagues so they can dominate games when they find they are playing against tougher competition.

While playing the game this weekend that is what I saw in 11 of the 14 games I played. I was completely disgusted with the gaming experience. Tankers, leavers, MK2 dash bots with MK2 weapons in silver league racking up 1+ million points. No thank you Pixonic.

I think you bring up a very good point regarding EA. If Pixonic sees they are losing their player base and promptly acts accordingly, they will have come to an understanding that they can have a fair matchmaking system and release weapons that have and advantage but not such an absolute advantage that they never lose. I think this is critical to their survival. The money may be rolling in now, but if they are losing their player base that is not sustainable as has been proven by a long list of now defunct games and game companies which are no longer around.

As you pointed out in your EA example, such a system is possible and they will still make money because those P2W weapons provide that extra firepower to give the P2W players a subtle advantage, but not so much that those without the weapons are constantly having a bad gaming experience.

Thanks for replying to my inquiry. The reason I asked is I had a theory on why some people say matchmaking is not a problem.

The theory is that the closer your weapons are to MK2 the less that matchmaking is an issue for you. Based upon my own experience with level 6 heavy robots, I can effective compete with robots/weapons 2 levels above mine. If they are at level 9, I seldom can overcome the difference in firepower when in a 1 on 1 situation.

I want to thank you for providing the information on your lineup.

FM9CUY - I want to discuss shocktrain for a moment. Without question, it is the most damaging weapon due to it's range and "riccochet" or "lightning" effect. The problem Pixonics has created is they have put out a weapon that is potentially many times more powerful than the next most powerful weapon. It would be hard enough to do matchmaking based upon hangar strength with the shotgun type weapons (any weapon where the closer you are to the target, the more damage you inflict). But here Pixonic has created a weapon that doesn't rely on distance or static damage amount per shot.

To give an illustration, say a bulg with 3 shocktrains can do 10,000 damage on 1 target. If I understand the way shocktrains work correctly, the damage goes:
1 Target Hit: 10,000
2 Target Hit: 15,000 (10,000 on the first and 50% of that on the second)
3 Target Hit: 17,500
4 Target Hit: 18,750
5 Target Hit: 19,375

Now tell me, what other weapon comes close to doing that type of damage? What type of firepower advantage does that give the player who has tanked down to silver level? (In my opinion, a shocktrain should have only been provided for 1-to-1 custom games.)
Looking at the comments, this is the reason people are saying Pixonic is ruining the game. I can speculate at to why. On your test server, everyone is playing nice, nobody's leaving or tanking, so the type of mismatches don't come up. What type of test server do you have if everyone is within a couple levels of each other in terms of weapon strength when on the production server that's not what is going on? In this article, https://warrobots.net/en/posts/88, although in the article there is a section "How public test servers usually work", it doesn't say anything about gameplay. Does Pixonic let players tank/leave/seal-club? If it doesn't emulate the real production environment, it's nothing more than a science experiments for new robots and weapons. If it does emulate the production environment with tankers and leavers, how can these weapons and robots be released to production when they are so over the top in terms of having a decided and overwhelming advantage in battle?

If you're not properly emulating the production environment with the test server, you're going to get horrible outcomes like Pixonic has created with the shocktrain. Is it possible this all comes back to matchmaking?

LoucustMaster117 gives a prime example of why shocktrain or matchmaking is not the number one issue to players at the "top of the food chain". LoucustMaster117 talks about the battle modes being the "better yet" problem to address.

There is a certain percentage of people who like all three modes, a certain percentage that only like 2 modes and a certain percentage who only like one mode. So maybe overall maybe 35% of the people like all three and the remainder only like 1 or 2 of the modes. Matchmaking can have severe impact on 80% of the players (assuming equal distribution of players having level 1 weapons up to level MK2 weapons and the top 2-3 levels being competitive with one another)

Horrible outcomes. That's the best way to describe what Pixonic has done with Matchmaking. I played 8 games with my boys yesterday. In all but 1 of the games there were not full sides. Usually 4-6 or 5-6. In two of the game, a player with a shocktrain racked up almost 1.5 million and 1.3 million in damage, destroying 12 and 15 robots. In the case of the 1.5 million, two of the players on the other team tanked and even with all our players we still lost. Of the other 2 players, neither of them scored more than 180k damage! What the h-e-double-l man? Why is it that when Pixonics thinks about balance, they think the best way to balance is through weapons
(https://warrobots.net/en/posts/114 "Balance Above All Else") and the continually nerfs and buffs.

In my prior comment I lamented about Pixonic lacking any sense of urgency. At ALL. The balance and matchmaking yields this to be true. I will focus my next (series of?) comments on Matchmaking (which will also discuss nerfing and buffing further).

Final note: congratulations, if you're still with us. You made it through your first week of moderating. :)

Regional servers....huh? Why? If you say because the original developer didn't know better, ok. But this has been a problem that was dealt with before Pixonic was developing Walking War Robots:

https://www.rockpapershotgun.com/2011/03/07/pikkotekks-1000-player-fps/

http://www.erlang-factory.com/conference/ErlangUserConference2010/speakers/DavidAlmroth

https://news.ycombinator.com/item?id=14120506

http://highscalability.com/blog/2014/10/13/how-league-of-legends-scaled-chat-to-70-million-players-it-t.html

https://www.reddit.com/r/starcitizen/comments/4xvty4/potentially_thousands_of_players_in_the_same_area/

http://www.erlang-factory.com/upload/presentations/395/ErlangandFirst-PersonShooters.pdf

Pixonic is looking for a workaround? Workaround has the implication that it's a temporary fix until a proper solution is implemented. There are a handful of hypervisor and virtual compute technologies that can unify distributed cluster compute ecosystems which are sustainable in the long term.

Fixing Matchmaking - 1: Foundation

In my opinion, matchmaking is an absolute disaster. It has created a number of problems that Pixonic has tried and failed to address...after 10 major release and several maintenance releases. Tanking/leaving, buffing/nerfing. I attribute all these issue with matchmaking. If Pixonix's primary goal was to eliminate the Gepard-Magnum seal-clubbing, they accomplished that goal. If Pixonic wanted to create even matches, they failed, miserably. In fact, the second goal of the new system was stated here (https://warrobots.net/en/posts/87)

"Second goal — better balancing. War Robots is a tricky game to balance by itself, but having the separation between arbitrary tiers where same robots may have very different performance complicates the process even more. Getting rid of it we will have a much better picture. You probably shouldn’t expect that your plain old Destrier will be able to take the Rhino head-on (come on, he wasn’t meant for that), but we shall catch annoying balance outliers much more reliably now."

What I find ironic about the new system is that you've replaced one arbitrary tier from the old systems with a new arbitrary tier (Leagues) in the new system, which can be manipulated by players.

What do I think is a better system? I have my opinion, as do most people. However my opinion is based purely on math. It does not take player skill into account. It does not take into the advantage one weapon has over another. It doesn't take win/loss percentage into account or strictly look at if your robots are heavy, medium or light. It does not use any artifical tier, rank or other method to determine the outcome.

For the illustration of why math is the only fair and equitable basis for a matchmaking system, I will paste a section of a comment I have from some time ago from another comment/post:

"HOW MATCHMAKING SHOULD WORK: A one on one battle of a light bot at level 12 with 2 level 12 weapons VS a heavy bot at 6 with 4 level 3 weapons.
Light Bot Damage Absorption - 80,000
Light Bot Damage Infliction - 140,000
Heavy Bot Damage Absorption - 140,000
Heavy Bot Damage Infliction - 80,000
The damage infliction is the total amount of damage the robots weapons can discharge during the entire 8 minute match. This needs to be an assumption you make for matchmaking purpose: that a single robot can last the entire match. Maybe it will, maybe it won't. But that is where player skill comes into play and who you team up with in the match. NEITHER of these (player skill or who you team up with) should be part of the matchmaking formula."

The basis of a fair and equitable matchmaking system should utilze a total score for each player's hangar. It can not be manipulated. There is no benefit to leaving or tanking.

Well, isn't that the basis of the old system. Sort of, but there were several things the old system got wrong? A nearly complete list is here:

http://warrobots.wikia.com/wiki/Matchmaking(old)
#1: "Find the hangar power of your strongest bot. Your average opponents will be within +- 100 of your score."

The old MM system did not calculate the score of each robot's potential damage or it's strength (how much damage it could take). This was bad because if your strategy was to quickly get one robot up to MK2 and leave the other at level 6 (or whatever) then you were going to be at a disadvantage the bigger the spread in the level from your strongest to weakest robot. The fix would have been to do a calculation for each robot and sum it to come up with the match score.

#2: A +/-10% deviation from a 50% win rate correlates to +/- 10% variation of hangar score.

I wouldn't use Win/Loss to give you a range of value. It's too random based upon who you were teamed up and has a negative bias against good players and positive bias against bad players (handicapping). Unless you want to handicap matches since theoritically over time the randomness of being teamed with good/bad players will even out.

#3: WP bots increase hangar score by 20%.

Shouldn't be used.

#4: MM will continue to reach further up and down the scale until a room is made.

Well, if you're going to do this anyway then why not just do incremental expansion of all players in the pool of those waiting for a game in 5,000 or 10,000 damage increments to the players cumulative damage score to find the matching team members? That would be the most equitable way to provide ranges for each player.

#5 Keep in mind that there is a 145 point difference between lights and mediums, and a 145 point difference between mediums and heavies. 140 points is equal to: 3 bot levels & 2 weapon levels, 1 bot level & 3 weapon levels, and 5 bot levels & 1 weapon level. As long as your lighter bots are one of these combinations above the next weight class, your hangar is considered even (all within 5-10 points of one another)
Again, these point differences should not be used. The illustration above how the light/medium/heavy issue could be dealt with equitably.

The article went on and talked about the 99 scales and the components for the scales. Again, no scales or artificial banding should be used. Math based upon the amount of damage each robot can dish out and take is the most equitable system.

To be sure, some weapons and robots have advantages over others, but over time, after enough battles, this will equal out (provided the range of damage from the most powerful to least powerful weapon are within reason, otherwise everyone will just use the most powerful weapon - hello shocktrain?). It comes down to statistics and math.
So how can these calculations be done quickly? If Pixonic would use lookup tables which map the amount of each weapon can do during an entire battle with each weapon level and the amount of damage each robot can take at each level (1-MK2) then these calculations would be very easy and very fast. These tables would reside on the Pixonic game servers.

Multiple hangars should not be an issue. For those with multiple hangars, players are queued to a map and select their hangar. Once the hangar is select the calculation takes place and MM works its magic.

What about wait time for a match to begin? I don't know the impact this system would have. Maybe almost none. But here's an equally important question, how long are people willing to wait to be put in a match that was balanced? This would be a good poll question. 0-15 seconds, 16-30 seconds, 31-45 seconds. 46-60 seconds, 61-90 seconds. Pixonics: Instead of making assumptions, engage the community and make your decision on the feedback before you put something into production. A good suggestion, and not just MM!

Benefits of a MM system based upon weapon and robot strength:

- There is no benefit to tanking/leaving. The next game, with the same hangar, your MM score will be the same.

- You'll see more variety in hangars and matches. You won't see the lineups dominated by heavy robots.

- You'll see all light, all medium and all heavy lineups all the way up to Masters League. Because of the way scoring will work, and because of actual damage and current league being out of the calculation,, it's possible for lineups with low strength robots to be as high as Masters League. For those who don't like the heavy robots because they are slow, they will now be able to compete in any league.

Next up...Weapons, Damage, Damage Spread, Nerfing & Buffing

...stay tuned.

Curious as to what coding platform WR was written. Do you know?

To continue....
" I can explicitly assure you that you have a lot of people in the general players' corner than you know who constantly provide feedback..."

I am more than a bit confused here. To say one is in a corner came from boxing. Seems to me there are 3 corners as I don't believe there to be a general players opinion on the matter: 1. People who like the current matchmaking. 2. People who are indifferent. 3. People who dislike the current matchmaking.

Please clarify.

I can appreciate your tapping on the brakes reply. Here's the problem. When a train engineer taps on the brakes he knows how fast he's going, how much he needs to slow down, how much brake pressure to apply based upon the load, and most importantly, the time he has. Time he has to slow down enough to avoid derailing while on the switch and time he has for getting to his destination. Pixonics problem is two fold with respect to these endless nerfs and buffs. It appears Pixonics doesn't don't know the load, how fast they are going or how much they need to slow down.

The "tonnes and tonnes of data" isn't being collected, analyzed correctly, or both, to make the appropriate changes. The second item is there is no sense of urgency in fixing any of these things. It's clear there is no internal policy on addressing issues quickly. Look at how long matchmaking has been a mess at the standout example.

It should be obvious from the comments here and on facebook people have had enough of the perpetual "tapping the brakes". It's not working.

I remember you said you won some equipment in one of the story giveaways, so I have a story for you.

Three moderators are in the Pixonics breakroom having a coffee and having a chat about their weekend adventures in domination (in the game and real life). A developer walks in with a worrying look on his face. "What's going on" Dev says Mod1.

"Well, I've been told to change matchmaking but I don't know where to begin. Say, what do you guys think should be done?".

Mod1 replies, "Well, I think the current system is fine. So why change it. I have all MK2 bots and I can game the system so I get down in the silver and bronze leagues and seal club guys with robots with only level 3 and 4 bots. I can get a lot of gold and rack up the beacons and kills, and a lot of silver any plow through these events in no time and collect all the prizes. It's great."

Mod2 replies, "Well, I've looked at the players comments and I would say that about 50% like the current system and 50% don't like it. Not sure that helps but that's where the community is at on it."

Mod3 replies, "Well, I look at the system and see that it really isn't fair to people who have weaker weapons. As Mod1 said, he can game the system so that he is playing people with weaker weapons. You guys know I haven't been playing the game that long and don't have anything better than level 7 heavy robots and can't get above league silver 1 with my current lineup. So I am constantly seeing MK2s and dashers that kill my robots in an instant. It really ruins the gaming experience".

Now which of those comments are most/least helpful? It should be obvious that Mod2's comment is the least helpful. The comment from Mod1 and Mod2 are equally valuable as they both convey the details from a gamers perspective.

When I ask what your opinion is I ask from your perspective as a gamer. I understand your role as a moderator, but I was asking to know what your position is because I was presuming that the opinions of the employees of Pixonic (as players) are incorporated in the decision making process. When you stated, "Even the bot itself has nerfed a bit according to the suggestions.", are those suggestions exclusive of Pixonic developers, testers and other employees who play the game?

I think much of the information Pixonic collects is lacking in detail and not actionable. Case in point. On facebook there was a survey on the 3.6 release. The poll asked if we liked or did not like the release. Nearly three of four players said they don't like the release! Ok, so how will you going to fix it? You can't, reliably. What I mean is you can use the Facebook API and scrape all the comments and categorize them and if they are sufficiently positive or negative, Pixonic can come up with a plan to address them. But that isn't really reliable. Reason being that most people will answer a survey but not provide a comment in the "comment" section of a survey. It would be much better to incorporate surveys giving the list of the new/changed features and distinct options (liked, didn't like, didn't care/neutral).

Now you have specific items which are actionable, the percentages for each (most/least liked items) and from that can determine the order in which the negative items can be prioritized. Simple. Efficient. More reliable.

Minor correction, Clarke's third law references technology, not science, such as the physics and properties of electricity. (If I recall correctly, he was referencing future technology and writing more as a science fiction writer, not as a physicist.) If Pixonics course is to create weapons and robots that unreasonably defy know physics, such as shooting and walking through what are impenetrable walls or other such science fiction nonsense I am done with this game.

The developers have already made a huge mess with a weapon such as shocktrain and I'll detail why in my response to Loucust117 which, what should be no surprise, also ties into matchmaking.

Fixing Matchmaking - 3: Shields

One of the items I forgot about mentioning is concerning shield type robots such as galahad or carnage.. To be honest, I forget if you get damage points for damaging/destroying shields or if you don't. I seem to recall you don't. If that is in fact how it works, I would recommend changing it.

The reasoning as well as the calculations are straightforward.

Since you are expending firepower to take down the shield you are in fact inflicting damage, but it is on the shield, not the robot itself. Whether the shield is physical or energy it shouldn't matter.

Therefore, the amount of damage a shield can take before being depleted/destroyed should figure into the cumulative damage score (recall that calculation consists of the amount of damage a robot can take plus the total potential amount of damage each weapon can inflict during the course of a full time battle).

Using the above logic as justification, I would include the amount of damage a shield sustains before it is destroyed in the formula for any matchmaking calculation.

For those who want the shocktrain to be a weapon that requires more skill to use, how about this:

If the robot with the shocktrain only hits one enemy robot because no other enemy robots are in range or line of sight of the initial target, the shock comes right back to the bot who fired the shocktrain and inflicts 50% damage.

Now that would be interesting to test on the test server. Very, very interesting.

FM9CUY?

To be sure, the potential for big damage points is there. However I believe that would rarely happen for a couple reasons.

1. I believe it would require the person with the shield to be more or less cooperating with the attacker to let it happen. It's not like someone with a carnage couldn't close in on any attacker to stop it.

2. To rack up the damage it would take a considerable amount of waiting for the shield to regenerate. With other opponents on the map I don't see someone having the luxury of waiting around for the regeneration to happen before another opponent targets their robot.

A question on the energy shields. Do they regenerate more slowly as they absorb more and more damage, to the point they don't regenerate any more or do they regenerate indefinitely?

I would like to add that if there are no other enemy robots are on the map there is no 50% return damage.

Huh? That's not how it works in Firefox. New comments show up at the TOP (closest to the article), until there are about 210 comments, and then the order reverses so the oldest comments are at the top and oldest at the bottom.

But wait, there's more.

After about 230 comments there is no chronological order - they are all over the place.

Something is messed up. Look at the order of the comments here and then the order of the comments on https://warrobots.net/en/posts/215

Why the Test Server Blows - Part Uno

I halfheartedly made the suggestion on shocktrain as it is very extreme. It go me thinking though about the way the test server is used, the data and the seemingly endless tweaks when rebalancing of weapons and robots. So yea, I'm going to complain a bit.

I've already expressed how I think the way the test server is useless. Maybe it not just the test server but the policy about rebalancing. As I recall, what you said is Pixonic likes to "tap the brakes" to make these adjustments. The justification was so the adjustment doesn't have a strong negative reaction from those impacted (I'll use shocktrain as an example as it's now going through the rebalancing process). What about the other side which really has a much broader reach - those who have a strong negative reaction during the entire time the weapon is being rebalanced who see little difference in the amount of damage they get hit with after a new release where the weapon has been rebalanced?
It seems to me that Pixonic would want to get to the desired result as quickly as possible to avoid negative feedback from either side in the shortest period of time. So I want to ask, would Pixonic agree or disagree with the statement that the current dragged out approach has much broader negative impact on the community? If Pixonic disagrees, I'd like to hear the reasoning.

Going on a related tangent, this brings me to the test server. Ugh! How useless it has proven to be. I already mentioned how it doesn't remotely simulate production server gameplay. Putting that aside, if Pixonic has massive amounts of data at their disposal for analysis, they either are analyzing it incorrectly or don't have the correct level of detail. I suppose there is a third option that in the end it is the developers or someone who is making a biased "gut call" decision because they don't think something will/will not have the desired outcome. In a way you kind of alluded to this in a previous reply to one of the comments (mine or another? don't recall).

So the question is why does it take so long to come to the desired outcome? Policy, data, analysis or something else?

Until I hear back from you, I will assume the problem is with the data and the analysis activites and will comment on that next...

Check that...Oldest at the top, newest at the bottom.

Why the Test Server Blows - Deux

It seems there is a huge disconnect between the outcomes on the test server and the outcomes of the changes once they hit the production environment. Aside from the matchmaking system allowing and rewarding people who tank and leave battles there is the question of the analysis and data being used and how decisions are being made.

I want to emphasize again that using individual's "gut feel" and personal experience as the basis for decisions will almost always yield undesired results because it will incorporate personal bias based upon their game experience. That will include their preference for weapons, strategy, skill, and so forth. It their bias is not mainstream and they have higher rank inside Pixonix in the decision making process then we get what we see today. Maybe this is happening, maybe it isn't, which is why I'll focus on the availability of data and analysis on the test server.

To do any type of meanful analysis on the test server, at a minimum Pixonic should have the following available for each battle:

1. Pilot level
a. Each Robot type and each Robot's level (1-MK2). You also need to have the amount of damage each robot can take before it is destroyed (it will become clear why damage a shield can withstand is important later).
b. Each weapon on each robot and each weapons level (1-MK2).
How much damage each "shot" can inflict. Shot is the lowest level of detail. Some weapons will require minimum-maximum if shotgun type weapon such as thunder

2. Battle statistics for each pilot
a. The opponent's robot that was hit (robot type and level)
b. Amount of damage each weapon on each robot inflicts (amount of implied "damage" inflicted to shields should also be included), physical or energy. Implied means you attribute the damage value of each "shot" which hits the target.
c. Distance between robots when damage was inflicted (both robots, received damage, inflicted damage)
d. Robots played and destroyed, robots played but not destroyed (and percentage of "health" at the end of the battle.)

It would also be helpful to have battle time information on the above to do some analysis of how long robots last in a battle before they are destroyed.

With the above you have the basic information that can be analyzed and provide meaningful statistics. How would I go about the analysis of the data? First I would want to know what my desired outcome is. Everyone complains this weapon or robot is over/underpowered, so I need some hard numbers to compare the performance of each robot and weapon. So you need to understand where the weapons and robots currently perform. What you need is a baseline.

I would first baseline the production system and then extrapolate that information to the test system since they don't line up in terms of how players act in each environment (tankers/leavers). Then I would do an actual baseline and see how the two lineup to see where the differences are at.

But how to I look at the data? Well, I would look at the performance of each weapon (aggregate level 1-MK2 and then look at the damage of each weapon at each individual level). I would also look at the performance of each robot by robot type (again at aggregate 1-MK2 and individual level). When I talk of performance for a robot, I talk about it in terms of how much damage each robot takes and dishes out for each battle.
But the above involves a huge amount of data and data points. How do I visualize it and not have to look at 10s of thousands (or more) of data in an excel spreadsheet. Here you need to use some of the statistical tools that allow you to visualize the data.

In general terms, the data should follow a normalized distribution chart (or graph) such as this:
http://www.bidproposalform.com/wp-content/uploads/2017/02/normal-distribution-chart-normalstandard.jpg

So that is how you should be viewing the data conceptually. How would you view, say, the data for each weapon so you can compare them.

Ideally you would want to look at the data points plot in a 3 dimensional type chart so you can see the distribution of each weapon's damage allowing you to compare where the mean and damage distribution is at in relation to one another. Unfortunately I couldn't quickly find any 3d distribution models on the internet, so here are a couple 2d examples to give the general idea:

https://upload.wikimedia.org/wikipedia/en/thumb/1/1c/EMG_Distribution_PDF.png/360px-EMG_Distribution_PDF.png
https://www.umass.edu/wsp/images/poisson3.gif

What I would expect to see are the peaks (mean) for each weapon being offset as in the diagrams. They almost certainly would not be in alignment (but if they were, it wouldn't necessarily be a bad thing), with the mean for the shocktrain and some of the other more powerful weapons being to the right (higher damage values).

You would also want to do this for each robot and do some analysis to look at those details. Finally you would want to look within each robot and weapon level (such as comparing the robots and weapons at level 1 to see how their means change in relationship to the cumulative mean for level 1-MK2 robots and weapons).

So, getting back to the test server and nerfing shocktrain.

How would I do it.

Going back to my example (having the shocktrain damage bounce back to you) was a bit extreme and is very situational, completely changing the dynamics of the weapon. So lets step back and look at the data and how the shocktrain works. Based upon how the shocktrain works, I will assume that the amount of damage a shocktrain can do has the broadest range of any weapons available due to the "train" effect. How would I counter/manage that?

I would make the change based upon an assumption or premise that shocktrains are able to make most of their damage due to the distance of their reach. So I would make a dramatic change on the test server and change the distance down to 250m or 300m.

Why this distance?

At 250m the shocktrain is well within the reach of all weapons, the minimum being the Orkan and a few others at 300m. By looking at the damage distribution you could confirm that the Orkan is one of the weapons that inflicts the most damage in a battle and is the closest competitor to the shocktrain (maybe its the taran, zeus, or even thunder, but that is the purpose of collecting the data and analyzing - so you're not guessing).

So you make the change, run a couple days of battles on the test server with this configuration and analyze. Is the outcome higher/lower or very close to what you intended to accomplish? You would look at the mean values of all the shocktrains and the mean values for all the other weapons and now the basis for either running another test (maybe increase the distance to 300m) to again measure the outcome.

If Pixonic wants to make the game more competitive, they should try to move the means of each weapon closer together and the spread of damage (lowest amount of damage to highest) tight.

You could then use that data to run simulations with the data from the production system to validate the changes lineup with what was seen on the test server, but that is another topic.

Final item worth noting is quality of data. As long as the current matchmaking system is in place Pixonic will need to deal with garbage data caused by the "tanker" and "leaver" and exclude that from the calculations for determining the mean and the damage calculations.

As I have said before, all of this. The foundation of the game and game experience; it all comes back to matchmaking and having a system that can not be manipulated and "gamed".

Hope this give you guys some ideas on what data is needed and how to better use analytics to reduce the amount of time required to balance robots and weapons.