Spotting shermans too hard?

[akd]

Oh noes, I've been sucked into the YouTube wurmholllllllllllllllleeeeeeeeeee....

[c3k]

I wouldn't have watched...but for the Tobias Funke comment. I admit I had no idea what they were singing about...the first four times I listened. Suddenly, midway through the fifth replay, I suddenly understood Finnish. I found my inner Laplander. Either that, or it's simply the beer...

[ZPB II]

Since the train derailed and is plunging down into the Offtopic Canyon, there's not much to do except silent horror and acceptance. It's 6am and sleep has yet to grace me.

I'm glad people got something out of it. I love camp and kitsch. The song and video is from 2010 but the featuring singer is a reasonably esteemed oldschool artist with a decades long track record.

I'd say some of the themes are hypocrisy, rehab, existential pain, guilt, admittance & acceptance, perhaps a man who feels his time has passed?

Here's me spending, uh, roughly 10 minutes hastily translating the lyrics. I am fairly certain it will be butchery but I guess it might get the point across:

Awake alone staring into the night

A restless feeling takes over

I spin, I scratch

I demand answers from the dark

I fear the walls talk to me

Silence comes at 4:20

Just as I got used to the voices

I knew time was running out

CHORUS:

When I heard about it, it was too late

Autumn took my senses away

What to expect from life

When your brain is pounding all night

When I heard about it, it was too late

Everything had gone with the leaves

What to expect from life

When in a year I get out of here

Walls are no longer walls

Days go by, my time stands still

In the waiting room I hear my name

An old man begins his cure once again

Hands covered in soil I made it through the autumn

Second stage with a pale face

From the pharmacy I will receive my salvation

A young girl hands the pills

Then chorus chorus some more chorus.

[Michael Emrys]

I really hate rap.

[sburke]

Screw spotting Shermans. Spotting what the hell this thread has become is bad enough.

[ZPB II]

Screw spotting Shermans. Spotting what the hell this thread has become is bad enough.

Spotting is random and unpredictable. So is the Internet. This thread is a demonstrator.

Makes perfect sense to me.

[BDW]

LOL. we'll i'm trying to get an armor only QB going with Darknight. I need a good QB ME map. I can't seem to get anything good from the built in maps. Any suggestions? anyone have a good map?? something large so we can really put this spotting to the test??

[ZPB II]

LOL. we'll i'm trying to get an armor only QB going with Darknight. I need a good QB ME map. I can't seem to get anything good from the built in maps. Any suggestions? anyone have a good map?? something large so we can really put this spotting to the test??

Well, I have one map lying around that I consider to be pretty finished in terms of terrain, the scenario it is for is still WIP.

It's, I think 1600x1200m, not very large but I made it so it has a sharp contrast, featuring both wide-open bits with very long LOS and small, claustrophobic bits with little valleys and forests.

I could quickly convert it into a QB ME since there are parts in the middle of the map that should work as good objectives and it has covered deployment for both sides.

If you are interested? It would be nice to get some feedback on the map. I can't promise it will be good, though. I did have a quick look at the map and I think the variety could make it useful for this purpose. It won't have a briefing or any of that polish, just the map and objectives.

(I forgot QBs never have briefings, silly me)

[BDW]

Zebulon, sounds good! shoot it over to me so I can have a look.

[ZPB II]

Send me your email address as a private message.

[dieseltaylor]

I'm far from saying ANY of these games are perfect.

I have tested many facets of the game. My tests refute some of your points. There are problems with spotting not working the way we'd expect.

You made some statements which are demonstrably wrong:

- Units in motion do spot worse than stationary units.

- Size of (or number in) a unit does affect spotting.

Yeah, spotting feels like it can be improved. However, feelings are not data. Facts will be more beneficial to narrowing down what should be tweaked.

Ken

Talking of facts why not tell us how your tests are structured? Or in fact release the test you use on spotting and find out if replicated on many machines that there is a weird bug liking a floating decimal that actually borks

something.

Also I think it should be explicitly stated under what version of the game testing or queries relate to. The danger of referring to early experiences or tests exists for all of us and that together with explicit confirmation of any tweaks in new game versions means we can know whether to junk earlier tests.

[ZPB II]

Talking of facts why not tell us how your tests are structured? Or in fact release the test you use on spotting and find out if replicated on many machines that there is a weird bug liking a floating decimal that actually borks

something.

Little do I know, but I will make the guess that a lot of the testing he does falls under the beta NDA.

I would also assume that the beta builds are incompatible with the retail build.

So basically we might be asking a guy doing a lot of charity work for us to do even more work by wriggling his way past legal and technical nuances.

Just thinking out loud here, no intention to slam anyone's opinion.

There is always the possibility of a nasty bug sneaking in a later version.

[dieseltaylor]

Good thinking ZPBII however I was thinking that they would be releasing what they HAD been testing rather than current skunk works.

Therefore users of the latest released iteration should be getting the same results as c3k AND if not that is interesting. As it stands we bodge up tests that may be open to criticism on structure.

[c3k]

dieseltaylor,

Good questions.

MOST of what I test is under various beta builds. The determination of what I test falls totally under my own purview. A lot of the tests I try to create lead nowhere. A lot the tests take a lot of work to get it on target.

A few examples:

- The Tiger gunner facing sideways: that was the result of of my testing. NOT that I knew the gunner was sideways. But the spotting was off. That was a test of accuracy for the 88L56 as mounted in the Tiger I compared to other weapons. That took every bit of spare time I had for over two weeks. I ran literally thousands of iterations.

It would be impossible for me to describe the exact methodology that led to my results. The final test had a 4km firing range with 20 or 25 lanes. I would run that test to get at least 200 results and run statistical analyses on EVERY shot. I had the Tiger (or other weapon) fire at a Sherman at 3km until the Sherman was destroyed. I tracked EVERY SHOT and MEASURED EVERY MISS. Tigers, Panthers, PzIV's, various Shermans ALL went through ALL the tests. (I still have nightmares...)

- The "Mortars don't miss after their first hit" issue. That got my attention. I asked for the test which a poster said they had. I was told it was easy to do myself. Fine. I created a new test. That didn't work. It took me the better part of a day to craft a test which would eliminate any outside influence and purely test mortar accuracy versus different targets. I tested different ranges, different azimuths, different ranges AND azimuths, different mortar calibers, different sides, different experience levels, different movement displacements, etc. And yes, I recorded EVERY SHOT and EVERY MISS DISTANCE (if any). That took two extra days.

- Spotting: I will NOT write an essay describing all my spotting tests. I have spoken up in various threads and given criticism where the tests obviously are flawed. Or, I have pointed out potential pitfalls if the tester hasn't shared his technique. It is best if everyone tries to create a solution on their own. I would hate to have a test not occur because the player thought, "oh, I've just been shown the ONE TRUE way to test, and mine was not that way."

- Back on the Sherman/PzIV spotting thread, I saw an obvious fallacy. A poster was comparing spotting by Shermans vs. PzIV's and comparing them to spotting by PzIV's vs. Shermans. Apples and oranges. If you want to compare how well unit A spots compared to unit B, you need them both to spot the SAME object.

- How many eyeballs does each unit have? If Unit A has 5 men and Unit B has 6, then you SHOULD get a skewed result.

- Binoculars?

- Scoped rifles?

- Relative sun angle?

- Experience?

- Terrain?

- Leadership?

- Intel?

Et cetera.

In short, whatever I start testing, no matter how much thought I apply beforehand, I _always_ find multiple shortcomings in the test setup I create. For the last 5 days I've been digging into something. I have only yesterday gotten the test scenario to work the way I had expected it to...last week. Now comes the data collection phase. It will take me about a week, I expect, to come up with around 1,000 iterations.

Why so many iterations? Well, look at the "a grenade knocked out my tank" thread. Outliers do occur. 1,000 iterations tend to damp out their influence. (As would as few as 100.)

And no, I won't reveal what I'm testing! That falls under the NDA and good manners. What I test as a beta goes through the beta peer review process. If I am wrong, then all I've done is wasted my time without creating chaos and false expectations in the outerboards. If I am right, then all credit goes to BFC for crafting the game and the software which reflects the expected results.

As to floating decimal points and whatnot, I'm running various flavors of CMx2, release and beta, on 4 machines. They are a mix of AMD cpu's, Intel cpu's, AMD video and Nvidia video. I finally got rid of my last XP setup (no longer needed), and am running W7/32, W7/64, and Vista64.

FWIW, I work full time, I am a serving member of the Air Force Reserves, and I'm trying to corral 2 teenagers while renovating my house and performing 2 engine swaps based on 3 engine blocks and full, down to the piston rings, rebuilds of the engines.

Tetchy? No. It's the refrain, in Finnish, of that horrible video going round and round in my head.

So, those are some facts coupled with some opinion.

Next time you're concerned that you'll wasted time with a test, post a screenshot and a savegame showing what you're doing. I'm sure you'll get plenty of input. Shrug off the noise and press on. Run a bunch of iterations and present it.

(Great, now I've used up my lunch-time with an overly wordy post instead of testing or playing. If ANYONE criticizes hoplite susceptibility to flanking elephant attacks, I blame dieseltaylor!)

With a to ensure you recognize this is in good humor.

Ken

[stikkypixie]

Sometimes I think it's a pity the engine is so coupled to the visual representation. If these tests could be somehow automated it would save brave people like you sooo much work

[dieseltaylor]

Thanks so much for your full answer. I am encouraged to know that you are on the case! : ) Your dedication is humbling.

My problem, and probably for many other "testers" is we do not know what is actually factored in by BF. I recall when the game out I was wondering whether the sun had any actual effect on spotting and fooling very slightly with it as I wrestled with the spotting anomalies. The spotting was - and I am generally referring to armour to armour - was definitely bizarre.

I can understand the idea that testing design can benefit from different inputs but am also wary of designing a slightly inferior wheel. Seeing how the "pros" do it would perhaps be educational as to the whys and why nots of a particular design of test and help "novices" in designing their own.

P.S. Whilst I cannot claim to be a statistics guru I have a long love affair with knowledge, and especially when made easy by being provided in graphic form and have all of Mr Tufte's books - first being bought in 1987.

http://en.wikipedia.org/wiki/Edward_Tufte

[LongLeftFlank]

Celebrity has made Tufte quite the shill though; I went (on the company dime) to one of his daylong seminars back in 2006 expecting an intimate classroom-sized audience at $600 a head (30). Wrong. Try 400 people crammed into the hotel's biggest ballroom. $240k for 7 hours of lecturing which mostly involved paging through his Holy Books (at least a boxed set wasincluded in the tuition). Minimal Q&A -- he basically punted any extended conversation to his website forum (which is quite worthwhile btw). To quote Sinatra: nice work if you can get it. In case you're wondering how he affords that sculpture garden on his Connecticut estate on a professor's salary.

He's the Suze Orman of the academic world. No disrespect to the relevance of his teachings though.

[Vanir Ausf B]

Great post! Having done a fair amount of testing myself -- 'though a small fraction of your output -- I know how time-consuming it is.

- Scoped rifles?

- Relative sun angle?

Do you know if either of these factors make a difference in spotting?

[dieseltaylor]

Sorry to hear that he seems to have become a moneygrubber. Given most humans can read many times faster than a lecturer speaks I have always thought those things poor value. Also as you cannot break off to process a particular concept without losing more talk it is even more redundant.

P.S. I do find it a trifle annoying that the media can go so overboard on various personalities even if they are just spouting commonsense - I had a quick look at Suze's wikipedia article and site. Unfortunately the US does seem to need a celebrity culture more than most nations and I am left wondering why that is. My gut feeling is the media realised or generated a need to populate a "global" village filled with a limited range of characters that we all[?] have in common. : (

P.P.S Which reminds me of the idiocy of people who need to see someone in the flesh or buy garments with names on at inflated prices. humans! Huh! .....

: )

[GreenAsJade]

Sometimes I think it's a pity the engine is so coupled to the visual representation. If these tests could be somehow automated it would save brave people like you sooo much work

Word.

What a sad sad connection that is. Of course, it's what a modern game programmer has to do to harness the power of the GPU .. if modern PC's came equipped with a dedicate Physics Processing Unit (PPU) then game programmers could decouple graphics from engine. Dream on...

GaJ

[Michael Emrys]

FWIW, I work full time, I am a serving member of the Air Force Reserves, and I'm trying to corral 2 teenagers while renovating my house and performing 2 engine swaps based on 3 engine blocks and full, down to the piston rings, rebuilds of the engines.

Ye gods! :eek:

The word 'workaholic' floats to mind, but I wouldn't get in your way for anything. Just want to say thanks for finding the time to put so much effort into the game. Wish I could buy you a pint of something useful. Would 10W30 do?

Michael

[c3k]

Thanks for the kind words, and the offer of the 10W30 (but I've got CASES in the garage Now, beer is a different matter...). FWIW, I did not mean to post a "look at what I do", but rather to focus on how much effort it takes to tease out results.

My work pales in comparison to the efforts put forth by others on the beta team, and even more so when compared to those on the alpha team.

Big group hug. Great. Now let's start machinegunning each other's men.

[sburke]

Great post! Having done a fair amount of testing myself -- 'though a small fraction of your output -- I know how time-consuming it is.

Do you know if either of these factors make a difference in spotting?

Can't speak directly to that, but way back last year while running tests on the protective effect of being behind a wall I came across an anamoly I could not explain and my first question to BF was - do you guys model targets being silhouetted? There answer was no (this was all on the public forum). The anamoly ended up being an effect that was occuring if your troops were placed in front of a wall and was corrected later.

It just goes to show that you may be seeing a symptom of something totally unrelated to what you think you are testing for so what you think needs fixing isn't at all the problem. :eek:

[sburke]

Thanks for the kind words, and the offer of the 10W30 (but I've got CASES in the garage Now, beer is a different matter...). FWIW, I did not mean to post a "look at what I do", but rather to focus on how much effort it takes to tease out results.

My work pales in comparison to the efforts put forth by others on the beta team, and even more so when compared to those on the alpha team.

Big group hug. Great. Now let's start machinegunning each other's men.

You make me utterly embarrassed to say I have beta tested squat.... My inferiority complex just went through the roof. I'll just go curl up in the corner in a fetal position.

[Vanir Ausf B]

Can't speak directly to that, but way back last year while running tests on the protective effect of being behind a wall I came across an anamoly I could not explain and my first question to BF was - do you guys model targets being silhouetted? There answer was no (this was all on the public forum). The anamoly ended up being an effect that was occuring if your troops were placed in front of a wall and was corrected later.

Yes, I remember that thread very well