6 replies to this topic

[SpiralRazor]

MustrumRidcully: Can you explain a bit about the patch release process?
I don't want to come off as overly critical or something here, but:
We've had at least two patches where missile flight paths were broken and one where missiles' splash damage interacted together with the hitbox change to make them overpowered, forcing hot fixes. Did the QA miss these issues, or did you find them too late to change anything for the build? You changed missile damage despite having flight paths and splash damage being off. I would expect the damage change would be tested, which would - unless it isn't done in game - necessitate testing the flight paths as well. How do these things fit togehter?

A: All of the above can happen and has. We have an extremely aggressive release schedule, one that sees a feature completed no more than 4 weeks ahead of going live. We’ve been working towards increasing this gap to allow for better testing and soak time. A new set of test servers will come online in the next 4 weeks. These will allow us to test major features and minor tuning adjustments at scale. Scale meaning player loads similar to our production servers. Between IGP and PGI we have finite resources and hardware. It’s almost impossible for us to replicate live conditions at scale. A lot of issues only rear their heads when we have 1,000s or 10,000s of players playing. With each month that goes by, we are able to improve our processes, and testing to improve quality and delivery of content.



With all due respect to the people who *actually* do there jobs at PGI/IGP, I find this answer to be a load of *********.

With just myself and a random group of people I may que with, I can find a broken game balance implement in under a few hours; A) From 1st hand practical application and from observed effect.

I dont test or make an analysis in a vacuum. Also, as a business manager, If something crosses my desk that has a line of figures that dont make sense, or even appear to not make reasonable sense....I look into it. Whoever was coding the missile damage/flight paths/component interactions should have caught it, and failing that, it would have been EASILY caught by a group of internal play testers within oh...I dunno, maybe an hour or two of play.

"Hey guys, I just SUPER tweaked missiles and changed multiple interactions at once, please have one team use some primary load Missile platforms and take some spotters and the other team needs to use a more standard set up as a control. Get back to me by the COB tomorrow pls."


Sorry, dont buy it because I have a brain.

[Deathlike]

Guys... PGI is the same group that deployed MGs with the wrong crit #s and with non-working crit chances when it was initially deployed.

Do you expect them to get it right this time around? Fat chance.

Whenever it is said they have an aggressive schedule, it's true, but it's aggressive in the wrong areas. There's getting it out first, and then there's getting it out RIGHT the first time. In the current industry where releasing games AND having a day 1 patch is the norm. This is no different. That is the problem PGI is not recognizing it an actual issue.

[Appogee]

A thought for you to consider...

Releasing content aggressively, and then patching it aggressively, leads to greater variety of player experience.

While some will rail against the frequent patches, some of which contain broken elements and create balance issues, these aggressive changes of weapon systems do create substantive changes in the way we have to play.

Variety is the spice of life.

So, let's be careful what we wish for. Frankly, I am close to walking away from the game due to the sheer monotony of these short battles repeated over and over, on a small number of maps and only two game modes. I actually look forward to patch day because there's a chance it will cause the game to become different enough to be interesting for a while.

Edited by Appogee, 02 June 2013 - 11:01 PM.

[Ralgas]

SpiralRazor, on 02 June 2013 - 07:47 PM, said:

1nt.



With all due respect to the people who *actually* do there jobs at PGI/IGP, I find this answer to be a load of *********.

With just myself and a random group of people I may que with, I can find a broken game balance implement in under a few hours; A) From 1st hand practical application and from observed effect.

I dont test or make an analysis in a vacuum. Also, as a business manager, If something crosses my desk that has a line of figures that dont make sense, or even appear to not make reasonable sense....I look into it. Whoever was coding the missile damage/flight paths/component interactions should have caught it, and failing that, it would have been EASILY caught by a group of internal play testers within oh...I dunno, maybe an hour or two of play.

"Hey guys, I just SUPER tweaked missiles and changed multiple interactions at once, please have one team use some primary load Missile platforms and take some spotters and the other team needs to use a more standard set up as a control. Get back to me by the COB tomorrow pls."


Sorry, dont buy it because I have a brain.

All im gonna say is there are plenty of other mmo style games that do the same, god half of wolk you didn't bother playing for 4 days after any sort of reasonable patch because it needed rolling restarts and tweaks to get back to playable and iron out "bugs" and "exploits" that cropped up

Planetside 2 atm has memory leaks by the look and crashes every 10 -20 min with gpu physics on.......

Par for the (unacceptable it seems ) course

[Vapor Trail]

LRMageddon I and II (AKA Singin' in the (LRM) Rain and Chicken Little Rides Again.)
'OneShot' SRMs (Single SRM6 takes down a Commando).
The initial implementation of Double Heat sinks (why is my mech cooler with singles than it is with as many doubles?).
Pulse-Laser heat too low (I figured the change was intended, and made the Large Pulse actually desirable).
Nova-glow mechs (I just shot a Catapult with a full alpha strike, and now I can't SEE).

All of those probably should have been caught by QA before the patch went live. No legit excuse on some of them. "At scale" hardly mattered for most of them. From memory, the community spotted most of them in a few hours after the patch went live. About the only one that slipped by was pulse-laser heat.

It wasn't a case of 50,000 people testing every permutation of gameplay. It was a handful of people who played the new patch, noted that 'X' felt wonky, said so on the forums, and had it snowball from there.

---

"We have an extremely aggressive release schedule, one that sees a feature completed no more than 4 weeks ahead of going live."

Is this stated correctly?

What I get from this: A feature is completed roughly twenty-eight business days before going live. Assuming a build containing thirty-two separate system changes (I doubt there's been a patch this size) that's almost a full working day (7 hours) of testing time per system change. Not seven man-hours, seven hours by the entire QA team. This does not seem to be an aggressive release schedule.

Or was the intended meaning more along the lines of "We have an extremely aggressive release schedule, one that sees a feature completed in no more than 4 weeks before going live."

Seems to me that this is more likely the case. But again, how is it ascertained that a feature is complete without testing it? The LRMageddon episodes argue that the feature either was not tested after code was finalized, or that the tested code was never properly implemented and the code that actually got signed off on as working was wonky in the extreme.

Stuff that's accidentally broken in the course of a patch, that was never intended to be changed, therefore was never tested, and therefore needs to be fixed I can see.
Stuff that was worked on, and apparently tested and signed off as working (and working properly), that is still broken, I can't see.

I'm still curious. What's the hold up on Collisions and Knockdown? They've been gone for for-ev-er. If/when it's reintroduced are we going to have a return to the days of Dragon-Bowling because one mech wasn't knockdown tested? Or will one of the mechs still teleport to sixteen different locations before the server finally decides where the mech is because some legacy code got implemented instead of the fixed version?

Edited by Vapor Trail, 03 June 2013 - 12:26 AM.

[El Bandito]

I really think PGI should let us have a community representative/s--especially on balance issue--and I am sure many people here would gladly nominate Mustrum for the position.

[El Bandito]

Thunderklaws, on 09 August 2013 - 04:52 PM, said:

We should, all the players of the game, get together and vote on a panel of say... 13 people, from all across the spectrum of gaming, then PGI could fly them all to Icel- er Canada and they could visit CCP- er PGI HQ and talk directly to the devs with concern to the game. We could call them the Council of Mechwarrior Management

No need to fly there, internet conference should be enough.