Revoluzzer

The baseline for 'confidence' is 0. That's when your threat is 100% volatile. There are only two types of players who will have a CV of 0 - new ones and dethreaters. For most other players the CV will probably hover in a "medium" position (assuming it goes from 0 to 1 that would mean ~0.5 for most players). Some players who perform very consistently will be close to the max value. How does the system know when your threat has to change from one colour to another? Simple, that's based on your glicko value. A range of values corresponds to a colour. A subset of values corresponds to a threat level within that colour (e.g. Silver 7). Since glicko works with fairly high numbers (afaik in the four digit range) I assume each threat level is composed of a relatively large range of glicko values (easily double digits). Now here's why I don't think the system needs to look at your past X matches' performance: To calculate threat you need only two ingredients: 1) The current threat level of everyone in the match, which defines the expected performance of everyone in the match; 2) the actual performance of every player in the match at the end. Did they meet their expectation? Keep their threat at the same level (rather unlikely to happen; at least some minimal movement is expected to happen at all times). Did they fail to meet their expectation? Adjust their threat accordingly. Why do we need a confidence value at all then? It acts as a stop-gap to reduce threat volatility and the constant change in threat levels (now hidden, of course) which would come with it. Why do we need this stop-gap? Because otherwise players would regularly have their threat-level change in unexpected ways (e.g. lose, but have it raised), which causes confusion and oftentimes anger. With the CV in effect minor spikes in performance (both up- and downwards) will be much less relevant and keep the system more stable and accurate. At no point is it required to look back at your past performance, because your current threat level / glicko value already represents how you performed in the past. That's how it got to where it is.

I don't think it really matters what the random bloom is caused by; it shouldn't exist in the first place. It's not a mechanic any player would expect from the gun and it was better off without it. It was a terrible addition in an attempt to fix the N-Tec.

Since we're on the topic whether most weapons can stun at all, I think stamina damage should be used more on some weapons to disable an opponents ability to sprint. Once your stamina goes below a certain threshold you can not sprint anymore. I think weapons like the SG-21 'Strife' should use mechanics like these to make them more interesting (i.e. does loads of health and stamina damage; kills slowly, but you can not run away).

Afaik it is not based on any recent matches. Not 10, not 50, not 100. As you play the system tries to figure out what performance it can expect from you. That means for each mission it estimates how many points you will achieve - or rather, how many points you will achieve relative to the other players in the mission. If your result is close to the estimate your confidence value (CV) will increase. This confidence system defines how much your threat level can and/or does change after a mission. With a high confidence value your threat level will change very little if you miss the system's estimate by a long shot. Instead the CV will take the brunt of the hit. This is the reason why de-threating usually takes a while to show effect, but then works quite rapidly. Fortunately it works rapidly in both directions then, because once your CV is low it will take a while to build up again. It's a shame the old blogs were taken down. But "Settle the Score" was apparently introduced in April 2013. So up to that point it was based on wins and losses. Another article here.

The snubby received an unnecessary nerf waaay back, having its damage lowered from 350 (same as RSA) to 300 (current SNR). Why did G1 do this? Because people figured out you could use some burst-damage weapons (mostly shotguns) in combination with the SNR to get very fast kills. Which just so happened to be what the SNR was designed for. Given its low rate of fire it was a high risk, high reward pick, but fun nonetheless. In other words the SNR isn't a great argument here, because it's currently not in its intended state (as far as I'm concerned). I agree with the general sentiment of Rifleman weapons gravitating around the jack-of-all-trades role. However the one true jack should be the STAR, because it is also a [i}master of none[/i]. The N-Tec is more effective than the STAR because it does flat out everything better. Other ARs do some things better, but others worse. The ATAC and ACER are intended for short range engagements and therefore should or do suffer at ranged combat. (The ATAC should have its damage drop off start at 40m, for example.) Suggesting in any way, shape or form to buff more ARs into the 0.70s TTK range is simply the wrong approach to weapon balance in my opinion. It muddles the entire lower range spectrum, which is already too messy. The COBR-A has no idea what it wants to be. Does it want to do ranged combat? The TTK certainly suggests it, but accuracy degradation and lack of recovery speak against it. It feels like an SR15 that took a page out of the CR762 playbook and misread it entirely.

I don't think other games handle this any different. If they did, the crosshairs would have to change size whenever the distance to the targeted area changes. Which I have only seen in less than a handful of games. It could be made more transparent, however, if weapon meshes (I assume this what @lgawe1994 referred to) would be select-able via drop-down menu, just like skins.

No, of course not. You get matched across servers. It doesn't matter in which server you are (Waterfront vs Financial does still matter, of course). I wouldn't be surprised if you don't actually load into a different instance after a while because the system will naturally distribute players of similar skill onto the same servers. Of course during low population times you will still get bad matchups or always fight the same group of opponents because no better/different match is available. Or you might get matched across regions (EU, US, RU etc.) which can provide even skills, but poor ping times.

Alright, let's go with the OBIR stuff first. I agree it's a powerful rifle. 990 health damage in two bursts can be pretty problematic. However they're not guaranteed, since you still need to actually land 6 shots. Can it do a lot? Yes-ish. It's of course great at long range, which it should be. It's okay at mid-range due to burst fire (corner-popping). It does more than tickle at sniper range. But below mid range you flat out need to be better than your opponent to beat them (this includes the use of "reverse" quick-switching). The N-Tec could kill you faster in CQC - purely on a technical level - than it takes to pull out most secondaries and fire them often enough to get a kill. And before that you still need to fire and hit an entire OBIR burst. The OBIR is harder to use than the N-Tec and still won't give you an edge over an equally skilled opponent anywhere from close to mid range. There is certainly no "enlightenment" when it comes to the power of the OBIR. It's simply a decent weapon within its niche. Balancing the N-Tec is not the be-all end-all, mind you. It's just a fix to one of the most egregious balancing issues. It will probably uncover more issues, albeit minor ones. People say now the Obeya might become the new "meta"? Entirely possible. Then that needs looking at. Others say people will instead switch to [insert a load of different weapons here]? That sounds like a more healthy overall balance to me. The general goal should be that more different weapons get used and players switch weapons depending on their environment. It's entirely common to see a whole team on a mission use only N-Tecs. Or N-Tecs and N-HVRs. It's just as common to see the top 6 players of each faction in Fight Club to use only N-Tecs. Or N-Tecs and N-HVRs. That's not good balance, that points to a major flaw. Now you look at all the weapons and say the N-Tec is the gold standard for weapon balance. And this might be a valid argument if all other weapons were utter trash and did not even perform well at their one, dedicated role. But on the one hand this isn't the case. Most weapons excel in their niche and only there - which is what they should do. Unlike the N-Tec, which not only excels at mid-range, but also does well at close and long range. Then on the other hand you need to be aware of gameplay design as a whole. There is a reason most pointman weapons gravitated towards a 0.70s TTK, while assault rifles targeted 0.75s. This creates a hard, technical limitation at which range assault rifles can not mechanically beat pointman weapons below a certain range. Of course random factors do play into this, but the favour always lends towards the pointman class. That's also why the rifleman class hovered around 1.00s and snipers around 1.75s. It simply means below a certain range these weapons need to rely on luck or the player being flat out better than the opponent. Why is all of this important? Here it comes: If you balance everything around the N-Tec, you need to buff a lot of weapons. A lot. And this will come with a plethora of issues. First, we are already aware that the N-Tec does a lot of things very well. In order to compete with that, other weapons need to do a lot of things very well. At which point you might still be able to have them handle a little different, but overall it doesn't matter all that much which weapon you actually use. They will all do well. I like to call this "The Call of Duty Problem"; all guns play more or less the same. Second, we are now also aware that weapon balance does not only concern weapons, but gameplay as a whole. If you need to buff a lot of weapons to get them on the same level as the N-Tec you will bring down the average TTK. Since the game world already struggles with the current average (i.e. oftentimes running from one corner to another is almost certain death), a lower average would be even worse (i.e. running from one corner to another is now certain death). I don't have a good name for this, but I like to compare it to "hardcore mode" in most games. It's hardly fun to die really quick (as far as I'm concerned), but its even worse when you have such long respawn and return times (i.e. getting back in combat) as in APB. In CoD "hardcore mode" kind of works, because you will spawn right back in combat anyway. Third, we are certainly also aware that buffing a lot of weapons is a lot of work, with room for a lot of error. So not only will you make the game worse as a whole, you will also open yourself up to creating even bigger issues while you are at it. Fight Club is not representative, because you will usually be low on health yourself or encounter enemies who are low on health. Furthermore you will usually encounter a large group of opponents or be supported by a large group of allies. The game is very much supposed to provide team-oriented, objective-based gameplay in varying environments. As such weapon balance needs to support this kind of gameplay, while Fight Club is merely an offshoot where all bets are off. Sure you can use it for additional data once you feel comfortable with the results you get from Action Districts. That's like taking a street car on the race track for additional data. But it shouldn't be your focus at all. I do agree that the entire test district situation is less than ideal, though. A more hands on approach in the live environment (i.e. small changes every few days) would probably work better.

Generally the rifleman class should have a slower TTK than the pointman class. It means they technically can not directly compete with pointman weapons, unless they get lucky. Of course there may be exceptions to that rule, like a dedicated close range assault rifle. Obviously this rule has not been adhered to in the past, which lead to such nonsense like the sub-0.7s TTK OCA. Buffing all the "weak" weapons instead of "nerfing" too powerful ones causes a power creep which is unhealthy to gameplay in general.

The crosshairs display your accuracy at 10m. Naturally shots can hit outside them beyond that range, because your accuracy is worse there.

Not a fan of this particular design, but you could expand this idea to ingame brand based items and sell those in similar sets.

The solution is already on its way, it's called 'phasing'. You will get matched across districts, rendering segregation irrelevant.

That doesn't make a lick of sense. Baylan Shipping heavily favours long range weapons (rifleman, marksman, sniper). The Asylum heavily favours short range weapons (rifleman, pointman). Coincidentally the N-Tec does well in both, which might lead you to believe it's perfectly balanced and everything else is broken. But it's not the weapons that are broken, it's your argument.

Those "outliers" aren't that rare in an environment where bad matchups and dethreating are commonplace. Alas, those will hopefully go away with phasing. Also confidence values aren't "wacky", they help alleviating the difference between good days and bad days a lot of players have. By having a confidence value the system prevents your threat from moving too much based on minor spikes in your performance. Since this confidence value has to be built up you will see new avatars threat levels bounce all over the place for a while, including odd situations like increasing their threat after losing against high threat opposition (another "extreme outlier"). Another nice feature of this confidence value mechanic is that dethreating is somewhat more difficult to do. In order to rapidly dethreat you have to drastically lower the confidence first. At this point you can, however, also rapidly increase your threat level again. And since the whole purpose of dethreating is to stomp weak opponents you keep your confidence low and your threat very mobile. If we didn't have this "wacky variable" it would be a lot easier to routinely keep your threat on a lower level than it's supposed to.

As a supposed "jack of all trades" it's surprising it does excel at anything at all. The STAR is already extremely easy to use and a great starter weapon, as far as I'm concerned. It doesn't require any deep knowledge of game mechanics to perform adequate. Since it is an actual jack of all trades, but master of none, it incentivises players to try other weapons. But they can always fall back to the STAR if they need something well rounded. Unless they try the N-Tec, of course. Then they have a weapon that is flat out better at everything the STAR does.

The more complicated bit is that you could also lose threat when you are in the top 50%, but perform worse than the system expected.

Making the rifleman class more varied while also buffing everything in it to N-Tec levels would have them take over the roles of pointman and marksman. As you correctly described the N-Tec can compete at both ends of its spectrum. If you want to take away from one end (i.e. create variety in the rifleman class) you need to buff on the other. So an assault rifle less capable of ranged combat than the N-Tec needs to be more potent in CQC. More potent? Shouldn't that be a SMG or shotgun then? But surely you want to retain some assault rifle character? So do you give it more range than an OCA, but the same TTK? Not less accuracy, though, because then it would be weaker than the N-Tec.

Match the ROF to the STAR. Should've had that from the beginning.

This is a fatal misconception. If you score a little above average, but against players of significantly lower threat, the system might also lower your threat. That is if your confidence-value is low enough. Because if it is not, the system might instead lower your confidence-value first.

That's probably because it performs really well and isn't very difficult to use. Unfortunately most proponents of the N-Tec use this argument to also say all other guns should be balanced around the N-Tec. Which would be necessary if you don't adjust it. But as @GhosT correctly laid out this would quite certainly lead to mostly samey gunplay. The N-Tec can still do everything from killing as fast as SMGs to reaching as far as designated rifles. Well it easily wins fights. Of course that's more fun than losing to an N-Tec user. Naturally people will try anything that beats them. And once they use the N-Tec those situations become less common.

Using JT rewards as an incentive to compete for higher threat levels is definitely a good idea. Since the current execution is lacking, I'd suggest to tie either the likelihood of receiving a reward or the amount to the threat level. For example Golds will always receive a reward (100% chance), Silvers oftentimes (75%), Bronzes sometimes (50%), Greens rarely (25%). Or at Gold threat you can get anywhere from 10 to 20 JT, at Silver 5 to 15, at Bronze 5 to 10 and and Green 1 to 5. Not sure if winning or losing should matter at all. Perhaps weigh towards the higher payout at victory and vice versa. Fundamentally tying payout to threat should incentivise enough to aim for victory. They are poorly designed and could be improved rather easily. Having the VIP spawn out of your reach before the enemy can get to him is just icing on the cake.

That one precisely. It's particularly problematic when you launch the game with headphones on and might have either not adjusted your volume yet (e.g. starting the game for the first time) or increased it before for some reason and forgot to change it back.

In a perfectly balanced roster there are no "meta" weapons. Of course this is nigh impossible to achieve. And sometimes you want to add some toys which are really specific in use, but not very great overall. That's okay, too. But no gun should be so well rounded it dominates several others in more than one situation. Or so powerful its individual impact is practically unmatched.

The current system is already Glicko-based. I don't see why they would use two different systems (ELO & Glicko), but Matt sure makes it sound like that - ELO for threat, Glicko for matchmaking. _________ Matt: "Scaleform isn't much better, it still runs fairly poorly out of the box, so we fixed it up. We are definitely doing all the new UI's, the new scoreboards the new district select screen, ... that is all on Scaleform." Whenever I read about Scaleform for UI elements, it was praised as easy to work with and performing really well (e.g. Borderlands). So is this an issue APB in particular has to deal with? Aren't current Scaleform elements still run through the Kismet-pipeline, too? I'd wager eliminating this obstacle should reduce the performance impact significantly? Kemp: "I'd love the old login screen because it was nice and simplistic, it got you into the game. It wasn't just kind of an eye sore, with exploding lights everywhere." It sure was an earsore, though. I feel like people ignore the nasty screeching sounds all the time. Getting rid of those, while not a top priority, should definitely be considered. Matt: "So a lot of the design work we're doing is just how do we make missions fun again." I wonder if you'll actually ever manage to do this. But if you can get a decent turf wars system going (also based on what you've learned with RIOT), which incorporates ram raiding & witnessing, you might find what APB has been missing; some decent, open world, player driven gameplay.

You don't have long matchmaking times when everyone in the district is at a similar threat level. Is this so hard to understand? It's a simple quality over quantity solution.