TR Forums

For a few people (including myself ) who are interested in participating in an upcoming beta test of this game this info might be useful:
http://blogs.battlefield.com/2013/09/th ... -bf4-beta/
Interesting points to note:
- The beta client is 64-bit only, so you must have 64-bit OS to run it (the final game will come both as 32-bit and 64-bit).
- The beta will introduce new additional anti-cheating engine: http://www.gameblocks.com. It is based on analyzing the in-game behavior and statistics and doesn't care if you actually run a cheating program on your PC, so regardless if you cheat or not - don't try to have a good aim (with a highly positive K:D ratio or too many headshots) and don't try to blindly predict the location of enemy's players (such as throwing the nades at the enemies who are not in your direct line-of-sight) or or you'll be marked as suspicious and banned/kicked

Well, that last thing is actually a very delicate problem to solve.

If they can't come to terms with normal human behavior predictions, and also edge cases, as in most maps would play in pre determined manner on certain game modes and will be prone to educated guesses, it could be a nasty surprise. Although today, pure statistical ratio based rules already are a nuisance since the basically don't work at all times.

But if they you can actually do a good heuristics engine that look at things it might work out just fine. I mean, AV/IDS and anti spam technology does that on a daily bases, if they can model a good baseline based on world class players, then put anything below that as ok, it would probably work out alright. 3:1 K/D in some rounds is really not a problem, 10:1, well, people flying in normal at some maps regularly get 50:1 so they really need to put together

Most decent players have probably been kicked from games at different times because they had a lucky streak or two. Getting a falling building down with a couple of squads in it are one of those cases, another is rush with c130 main gun on some maps where the defensive spawns are bunched up. Last but not least would be a lucky flank with a lmg that can take out a ton of people in short order.

Sweet so it is actually 64bit. I was wondering how they could be recommending 6GB of ram with what could have been just a 32bit game. Now to see how they put those extra resources to work.

Aphasia wrote:

But if they you can actually do a good heuristics engine that look at things it might work out just fine. I mean, AV/IDS and anti spam technology does that on a daily bases, if they can model a good baseline based on world class players, then put anything below that as ok, it would probably work out alright. 3:1 K/D in some rounds is really not a problem, 10:1, well, people flying in normal at some maps regularly get 50:1 so they really need to put together

10:1 is not really that hard to pull off every now and then. Most BF3 players are pretty bad.

A consistent 10:1 global KDR would be pretty special though...

Yeah a global 10:1 is probably next to impossible if you actually do any infantry at all. 3:1, probably not to hard to muster. I probably keep a 2:1 up to a 4:1 K/D nowdays, but considering much of the life time was spent at much lower K/D, it wont rise above a global 1.5:1 anytime soon.

Aphasia wrote:

Most decent players have probably been kicked from games at different times because they had a lucky streak or two.

Heh, I am a pretty average player (with less than 2:1 K/D ratio) but even I was killed by admin a few times and kicked once and got numerous accusations from in-game chat, and all I was doing is using Saiga with buckshot on CQ maps while trying to constantly move around instead of camping in single area and going up to 8:1 K/D ratio in some matches... I also played on some BF3 servers with FairFight already running and it did nothing to a very obvious cheaters (one of them was using a known exploit on Operation Metro map to get into "invisible" ceiling area, which, ridiculously enough, is still not fixed - luckily the admin finally appeared and manually kicked that person). So the whole "statistical analysis" method is still completely flawed.

The ONLY thing that will effectively combat cheaters is a conventional signature-based client-side cheat detection app, outsourced to one of the current antimalware companies such as Symantec/Softwin/F-Secure/Kaspersky Labs (so EA won't get blamed again that they are "stealing personal data" ). And, just like the current antimalware/antivirus programs, such anti-cheat tool MUST have a suspicious file submission feature, both manual (so gamers themselves can upload the existing cheat packages along with detailed descriptions) and automatic (if the program detects a suspicious activity by unknown threat), so the developers can analyze them and implement their detection into next signature update (which MUST come as rapidly as with current antimalware/antivirus programs). EA just does not want to spend money on more effective but (obviously) more expensive solutions...

Just FYI - I've been playing a few rounds of BF4 today with EVGA Precision's logging running... At the max settings and 1080p the beta build seem to utilize up to 2.4GB of video RAM, so I guess EA were right to recommend the card with at least a 3GB of video RAM...

It has awesome threading too, it uses 92% of my 1090T when on full settings, which is a new record for a game. I still get like 30 on full and 45 fps on lowest settings at 1080p however, even with a 7970....

JohnC wrote:

Just FYI - I've been playing a few rounds of BF4 today with EVGA Precision's logging running... At the max settings and 1080p the beta build seem to utilize up to 2.4GB of video RAM, so I guess EA were right to recommend the card with at least a 3GB of video RAM...

Rolling down to 2xMSAA should fix that, but good to know.

What CPU / GPU are you running and how's your frame rate?

Ari Atari wrote:

It has awesome threading too.

Yea, I noticed it too... It seem to equally utilize all of the cores, even the virtual ones (hyperthreading).

Prestige Worldwide wrote:

JohnC wrote:

Just FYI - I've been playing a few rounds of BF4 today with EVGA Precision's logging running... At the max settings and 1080p the beta build seem to utilize up to 2.4GB of video RAM, so I guess EA were right to recommend the card with at least a 3GB of video RAM...

Rolling down to 2xMSAA should fix that, but good to know.

What CPU / GPU are you running and how's your frame rate?

I didn't measure FPS, but the gameplay was pretty smooth. I have i7-2600 and Titan SC.

B.t.w, if you guys interested in seeing framerates with less powerful hardware - you can do that right now, on a (shameless plug!) http://www.twitch.tv/lirik stream (he has hardware specs right on his stream page).

Thanks for the reminder.

Origin is not as bad as I thought. I ordered it via the web at work. Logged into my home gaming rig and it was ready to download immediately.

I was also please to see that Nvidia prompted me to install their new BF4 Beta drivers (331.40).

JohnC wrote:

Just FYI - I've been playing a few rounds of BF4 today with EVGA Precision's logging running... At the max settings and 1080p the beta build seem to utilize up to 2.4GB of video RAM, so I guess EA were right to recommend the card with at least a 3GB of video RAM...

I'm going to have to look hard at this tonight- will have to see what it's going to take to keep the game reasonably in check on my 2GB cards. I'm not really interested in upgrading until I have a 4k monitor to use, and know what that's going to take... but I'll let y'all know how it works with a pair of 2GB GTX670's at 1600p. Shouldn't be too bad .

End User wrote:

Thanks for the reminder.

Origin is not as bad as I thought. I ordered it via the web at work. Logged into my home gaming rig and it was ready to download immediately.

I was also please to see that Nvidia prompted me to install their new BF4 Beta drivers (331.40).

Origin can be quirky, but it's still pretty solid overall; flexible enough to get the job done, while not being stupidly over-complicated.

I did run into an issue recently where it was telling me that my BF3 install was corrupt; I had to start Origin with Administrator privileges and run the repair, which brought it back.

But I also keep a BF3 install on a cheapo Verbatim SATA2 SSD that I shoved into a USB3 enclosure, and it works surprisingly well. And if Origin is started without the SSD attached, I just have to attach the SSD and restart Origin; don't even have to reload Battlelog. Also, that's accomplished through a symbolic link; Origin believes that Battlefield is still located on the local drive .

Works okay on a Radeon 6870 (on the new beta drivers); I'm getting mostly 30fps-ish on bare-minimum settings in 1080p. Also having problems with servers randomly disconnecting or alt tabbing causing crashes. I may try it again later.

Edit: Got better after reloading game so that lower res textures applied, but still stutters a lot. Phenom II x4 955 here.

I do not have issues with crashes but there are noticeable bugs, such as sometimes when you are killed you are not being put back into default "respawn map view". Plus there seems to be a problem with elevators (they either stop teleporting you or you randomly die inside of them) and hit detection (sometimes the game does not detect hits at all or sometimes you can be killed by bullets even if you're hiding behind a solid object)...

B.t.w, if you want to get a good "stress test" to your hardware - wait for the skyscraper collapse (your have to play in conquest mode) and just get somewhere near when it starts to fall down.

i7 3820 @ 4.4 GHz and EVGA GTX 670 FTW boosting in the 1150-1215 range...

60-80 fps, quite smooth overall with most settings to High and Ultra textures.

Deferred AA off, Post AA High.

I was worried I might have to yank my 2nd GTX 670 from my living room PC to enjoy the beta, given how buggy the BF3 beta was. So far it's not bad, I prefer Domination to Conquest.

JohnC wrote:

I do not have issues with crashes but there are noticeable bugs,

I keep getting what I think are PunkBuster crashes that throw me out to windowed mode when I launch the game.

JohnC wrote:

B.t.w, if you want to get a good "stress test" to your hardware - wait for the skyscraper collapse (your have to play in conquest mode) and just get somewhere near when it starts to fall down.

I'll look out for that. Thanks.

Prestige Worldwide wrote:

60-80 fps, quite smooth overall with most settings to High and Ultra textures.

Deferred AA off, Post AA High.

I was worried I might have to yank my 2nd GTX 670 from my living room PC to enjoy the beta

While BF4 looks great it does not appear to have much in the way of environmental fine detail. I started at Ultra and then dropped to High without any noticeable visible difference. My sessions were far to brief to really know for sure so I'm going to do some more testing tonight.

I ran the beta last night with a 3770K and a 770. Ran everything on ultra at 1920x1200, and it seemed to run perfectly smooth. I didn't measure FPS or anything (the render.drawfps function seems to be missing) but maybe I'll open up FRAPS this evening and see what it's doing. I should also see how it's loading up my processor. Maybe just maybe the 3770K upgrade will finally pay off over the 3570K I had.

First impressions of the game were good. Obviously a few things need fine tuning, especially graphics wise. I liked how 'smooth' everything seemed compared to BF3. Movements just seemed more natural and fluid. What impressed me most? When I hopped in a tank, I could hear the turret motor whining as I whipped the turret around. It was cool!

Great_Big_Abyss wrote:

I ran the beta last night with a 3770K and a 770. Ran everything on ultra at 1920x1200, and it seemed to run perfectly smooth. I didn't measure FPS or anything (the render.drawfps function seems to be missing) but maybe I'll open up FRAPS this evening and see what it's doing. I should also see how it's loading up my processor. Maybe just maybe the 3770K upgrade will finally pay off over the 3570K I had.

First impressions of the game were good. Obviously a few things need fine tuning, especially graphics wise. I liked how 'smooth' everything seemed compared to BF3. Movements just seemed more natural and fluid. What impressed me most? When I hopped in a tank, I could hear the turret motor whining as I whipped the turret around. It was cool!

In BF4, the command is now perfoverlay.drawfps.

Start to type "perfoverlay" in the console and you will see a few commands. Perfoverlay.drawgraph 1 is useful to identify CPU or GPU bottlenecks.

My CPU usage was 70% in BF4 across 8 threads on i7 3820 @ 4.4 GHz.

The quality of the visuals don't seem to be much different, just a different color post-process to make it seem different and shiny.

I agree that overall it feels smoother and more natural than BF3. I feel like the gunplay has been refined and polished compared to BF3. Although I feel like all of the n00b tube grenade launchers that most classes have is kind of unnecessary and dumbs down the gameplay, at least the damage on the grenade launchers seems to be nerfed compared to BFBC2 and BF3. However, there is still a full auto grenade launcher that is very annoying and OP.

I like it so far but my Q9450 is choking under the load. Too bad I have no room in my budget for a new computer =/

Prestige Worldwide wrote:

In BF4, the command is now perfoverlay.drawfps.

Start to type "perfoverlay" in the console and you will see a few commands. Perfoverlay.drawgraph 1 is useful to identify CPU or GPU bottlenecks.

Hey, that's great, thanks. I knew someone would come to my rescue with that info.

Prestige Worldwide wrote:

i7 3820 @ 4.4 GHz and EVGA GTX 670 FTW boosting in the 1150-1215 range...

60-80 fps, quite smooth overall with most settings to High and Ultra textures.

Deferred AA off, Post AA High.

I was worried I might have to yank my 2nd GTX 670 from my living room PC to enjoy the beta, given how buggy the BF3 beta was. So far it's not bad, I prefer Domination to Conquest.

Good numbers, but at what resolution / refresh, bro? E.g. 2560x1440@60Hz

Touché. I may have well meant 640x480 without specifying resolution.

The game defaulted to 1080 @ 144hz, I set it to 1080 @ 120hz / 2d lightboost once I noticed it.

edit: Put everything to Ultra, Deferred AA OFF, AA Post HIGH, 1080/120 and frame rate seemed unchanged. Still in the 60-80 range most of the time, going to much higher frame rates indoors and in situations like climbing ladders and looking upwards.

Seems a lot smoother than BF3 overall despite me having higher average FPS in BF3. BF3 had a few random stutters here and there. BF4 seems more polished and refined.

Most of my settings (other than MSAA) are maxed out and it looks like complete garbage for some reason. I think even Bad Company 2 had better graphics, let alone BF3. They've somehow taken a huge step back in terms of the visuals. It also runs a lot worse than BF3 for me (half the framerate and noticeable less smooth).

Is this what the game looks like for everyone? http://i.imgur.com/F57kNNh.jpg
I feel like I'm playing an xbox360 version or something.

It's still a beta build, with some of the features disabled, so yea, it does not look very impressive... Plus there is a lot of dust on this map (especially after skyscraper collapse) which also makes everything look more dull.

Oh man... That was fast

JohnC wrote:

Oh man... That was fast

(image snipped from quote)

I've almost considered suffering the pain of leaving the Master Race for BF4 because of this happening in BF3 so much. You don't see it, but it's so obvious that everyone getting a kill on you has some strange sense of situational awareness that you don't. Seeing it in picture explains it perfectly though. (And getting kills on these guys would be so satisfying, if only I knew for sure who they were!)

Yea... Sad thing is, EA still continues to rely on PunkBuster as a client-side protection, which is always too slow to react. Come on now, even my Kaspersky Antivirus was able to detect this "certain program" as "potential threat" for a very long time while PB does absolutely nothing except trying to occasionally take a screenshot... Which is actually useless because this "certain program" is able to mask the extra information and produce a perfectly "clean" screenshot.

I'm not surprised that the cheat surfaced now, because much of the game is probably reliant on the same base as BF3, with tons of refinements to things. But not reusing the part of the codebase that worked would've been wasteful.

But Punkbuster isn't a anti-cheat in the way that an AV/Anti malware is a protecting you. I wish it were, but it seems to be that Punkbuster observe, sending data to their end, after which you get banned server side. And IIRC they don't ban immediately too as not to give the cheat-makers a way to identify what they picked up. The problem with a client side detection is still that it is client side, and as long as people are voluntarily installing software, it's very hard to suppress it. It just makes it into a leap-frog contest of detect and evade. But punkbuster isn't nearly as good of a product as it should be, especially if they are supposed to take a screenshots and don't hook it low enough so something is in between it and where it captures the screen buffer from. On the other hand, cheats don't have to play nice with the system in the same way so they can probably take a few shortcuts that would be unacceptable to normal software.

That said, we still have no knowledge on how the new server side anti cheat will work after they have tuned it. But they stopped trusting Punkbuster as their sole thing. Which is good.

ps. love how they erased the names on the squad list, but left the squad names in the screenshots up. Sifting through a few battlereports with that combination of squad should easily narrow down the list of possible suspects to just a few. ds.