flight model accuracy?

[GOAT-ACEOFACES]

Will BoS provide any means of checking the accuracy of the flight models?

 

For example

 

1. IL2 has the DeviceLink option that allows the logging of some flight data.
    
2. CoD has the C# option that allows the logging of flight and other data.

 

• As noted in the "What we know so far" thread, Jason said DeviceLink will not be available at launch. But as most of you know, the true intent of DeviceLink was to provide home made cockpit makers a means to communicate with the program to not only drive cockpit gages, but receive inputs from the cockpit (switches, joysticks, etc). That is to say, using DeviceLink to log flight data was not what it was intended for.
• Also note that early on in the IL2 development Oleg made available the in game performance and some flight model data to a few members who in turn developed IL2Compare.

 

In summary my question to 777 is do they plan on providing 'something' that will show us the implemented performance (ala IL2 IL2Compare) and/or provide an ability to log in flight data (ala IL2 DeviceLink or CoD C#) both of which could be compared to real world data.

 

Thanks in advance

 

[BeastyBaiter]

Probably not but flight testing is easy enough to do. Find a test report on the plane in question and then duplicate their method precisely (RPM, radiator settings, fuel load and so on). The FM's in RoF rarely change and I'm sure the same will be true here. The game will only have 10 planes upon official release so it isn't like there are a huge number of planes to test.

[AcesHigh]

I had heard or read somewhere some of the team had actually taken up some of the ww2 planes and simulated some dog fighting.   True or not I can't recall just what I heard or read somewhere.  Seems to me if that's the case they are taking the flight model pretty seriously.

[DD_bongodriver]

I really doubt anybody handed the keys to extremely rare and priceless warbirds to some random game developers to go play with, what actually happened is they had a day out dog fighting in Yak-52's.

[Volkoff]

I had heard or read somewhere some of the team had actually taken up some of the ww2 planes and simulated some dog fighting.   True or not I can't recall just what I heard or read somewhere.  Seems to me if that's the case they are taking the flight model pretty seriously.

 

I can not recall any member of the development team representing that he or she had the opportunity to take up actual WW2 planes and simulate dogfighting, but I suspect that the source of the info from your recollection is somehow related to the publishing of this video:  

 

http://www.youtube.com/watch?v=d59wI2Uf6oA

 

Perhaps a community member on some board miscommunicated while describing the flight depicted in the video and you heard the faulty version of what actually occurred in the video. That happens to me more times than I would like. I end up thinking I have good info but instead I end up with a bad case of the Internet version of the game of telephone. The Internet can get information all over the World lighting fast. Too bad we can't put any confidence in most of that info!

 

 MJ

Edited October 24, 2013 by =69.GIAP=MIKHA

[AcesHigh]

I think bongodriver might be right they were not authenticate ww2 planes.   I can't remember where I read or saw it but they some simulated dog fighting in some aircraft as well as visited historic museums to study the planes in developement.   Maybe I dreamt it!

[LukeFF]

http://www.youtube.com/watch?v=d59wI2Uf6oA

[=AVG=Zombie]

Cool vid, They need guns

[AcesHigh]

Thank you for keeping me halfway sane!   Well OK maybe they were not BF 109's but they did fly in a single engine prop plane.   At 6:55 sure looks awfully similar to some of the screens I'm seeing in the vids

Edited October 24, 2013 by AcesHigh

[GOAT-ACEOFACES]

Probably not

Based on the fact that RoF hasn't provided any method (that I know of) I would have to agree. I was just hoping that with the merger, some of the things that 1C did, like DeviceLink and C# or providing the ingame data (ala IL2Comp) would make its way into 777s BoS

 

but flight testing is easy enough to do.

Disagree, it is easy to say, and easy in theory, but when it comes down to it, it is a very time consuming process. Which is why most don't do it at all or take the time to do it right. Especially with the DeviceLink and C# methods that require someone to actually fly the plane and collect data. Based on my decades of flight model testing (1992 red baron to now) most of the so called errors in the flight model are actually errors in the users test methods.

 

Find a test report on the plane in question and then duplicate their method precisely (RPM, radiator settings, fuel load and so on).

As noted above, easy to say, but in practice very hard to do.. For most people! A statement by me that is based on reviewing hundreds of flight tests done by self proclaimed experts. Where, as noted above, most of the claims of an FM error were actually errors in the way the test was performed. You would be surprised how many self proclaimed experts who claimed to have found a 'speed' bug in a flight model didn't even know the different between TAS and IAS.

 

I had heard or read somewhere some of the team had actually taken up some of the ww2 planes and simulated some dog fighting. True or not I can't recall just what I heard or read somewhere. Seems to me if that's the case they are taking the flight model pretty seriously.

Even if they did, data collected from a mock dog fight will tell you more about the differences in the pilots skill than it will about the accuracy of the flight models. There is a reason they had a specific set of test to perform in WWII to test the performance of a plane. Edited October 24, 2013 by ACEOFACES

[LLv34_Flanker]

S!

 

 AoA, get the Premium pre-order and jump in to testing right after 11th November I bet there will a lot of FM debates here after that

[GOAT-ACEOFACES]

AoA, get the Premium pre-order and jump in to testing right after 11th November

Got it.. have had it for some time now!

 

I bet there will a lot of FM debates here after that

Heck they are allready debaiting it!

 

Which is the reason for my post.. What is 777 going to provide us in the way of methods to check the accuracy of the flight models.. So we actully have something to debate about other than 'feelings'

[AcesHigh]

^ I've got a BF 109 and ju 87 parked in the garage if you want to take them up for a spin.  You have to pay for the gas though.  They are RC airplanes but never the less we could use them to test our flight models.

 

http://www.killerplanes.com/sites/default/files/imagecache/product_full/Dynam_Me109_PAINTED_14.jpg

Edited October 24, 2013 by AcesHigh

[HeavyCavalrySgt]

I am impressed at your commitment to duplicating power to weight ratios and control surface areas!

[BeastyBaiter]

Based on the fact that RoF hasn't provided any method (that I know of) I would have to agree. I was just hoping that with the merger, some of the things that 1C did, like DeviceLink and C# or providing the ingame data (ala IL2Comp) would make its way into 777s BoS

 

Disagree, it is easy to say, and easy in theory, but when it comes down to it, it is a very time consuming process. Which is why most don't do it at all or take the time to do it right. Especially with the DeviceLink and C# methods that require someone to actually fly the plane and collect data. Based on my decades of flight model testing (1992 red baron to now) most of the so called errors in the flight model are actually errors in the users test methods.

 

As noted above, easy to say, but in practice very hard to do.. For most people! A statement by me that is based on reviewing hundreds of flight tests done by self proclaimed experts. Where, as noted above, most of the claims of an FM error were actually errors in the way the test was performed. You would be surprised how many self proclaimed experts who claimed to have found a 'speed' bug in a flight model didn't even know the different between TAS and IAS.

 

Even if they did, data collected from a mock dog fight will tell you more about the differences in the pilots skill than it will about the accuracy of the flight models. There is a reason they had a specific set of test to perform in WWII to test the performance of a plane.

 

It's difficult at first but once you get the hang of it it's actually pretty quick (about an hour per plane). Honestly the hardest part is finding detailed original test reports to compare to. Basic performance docs are common and easy to find, but a technical report that details the methods, margin of error and gives detailed performance at regular intervals (1km is a good step size) from the deck to the service ceiling is a tough find.

 

Provided you have such things in hand, the quickest way is to test climb rate first and then test top speed on the way down. Reason being fuel load has minimal impact on top speed but can have a large impact on climb rate. It also removes most of the acceleration from figuring out top speed since you are actually diving to the desired altitude. So instead of trying to reach the top speed, you slow down just below it and then accelerate back up. Only takes 30s to a minute once at the desired altitude. Obviously a control test is required to verify that your top speeds are accurate with this method. To do that you simply repeat the low altitude top speed test after restarting with full fuel (or whatever the test report lists). It should be pretty close and tell you the margin of error.

 

I've done this a few times over at WT, and why yes, yes I did prove conclusively some of their planes are absurdly off the mark.   I intend to perform these tests on both the Bf-109F4 and the LaGG-3 in a few weeks when BoS beta is out for us. It will likely be one of the first things I do since I doubt it will contain much content.

[GOAT-ACEOFACES]

It's difficult at first but once you get the hang of it it's actually pretty quick (about an hour per plane).

It all depends.. The test itself is not the hard part, doing the research on how the tests were done back than along with the conditions of the test is the hard part. For example, some ROC testing was done from the min the plane started rolling on the runway, while others started the test once the plane was airborne. Something as simple as that can skew the TTC results by a min or so. But yes, for those who are use to it (read got the hang of it) it is easier to do than those who have only just begun or only done it a hand full of times.

 

Honestly the hardest part is finding detailed original test reports to compare to. Basic performance docs are common and easy to find, but a technical report that details the methods, margin of error and gives detailed performance at regular intervals (1km is a good step size) from the deck to the service ceiling is a tough find.

Agreed, and that if what I was and am referring to when I say you have to do the research.

 

Provided you have such things in hand, the quickest way is to test climb rate first and then test top speed on the way down. Reason being fuel load has minimal impact on top speed but can have a large impact on climb rate.

I typically break the two into two separate tests. For one, that is the way they did it, and two makes it cleaner with regards to data logging for later comparisons to the real world data.

 

It also removes most of the acceleration from figuring out top speed since you are actually diving to the desired altitude. So instead of trying to reach the top speed, you slow down just below it and then accelerate back up. Only takes 30s to a minute once at the desired altitude.

Problem with doing it that way is you will loose the acc data, which is useful for plotting Ps curves. That and that is not the way they did it back than, thus I prefer do it the way they did it. That and it removes the chance of anyone trying to poke holes in your results by pointing out you didn't do it the way it was done in real life. Note, I agree with you here, in that I know the fuel load does not have a big impact on top speeds, but, as noted I like the collect the acc data for Ps curve plots and I like to avoid the nit wits that try and say the method I used was wrong and thus the results are wrong.

 

Obviously a control test is required to verify that your top speeds are accurate with this method. To do that you simply repeat the low altitude top speed test after restarting with full fuel (or whatever the test report lists). It should be pretty close and tell you the margin of error.

That or just do it the way they did it from the start.. It does not take that much longer to climb from one alt to the next.

 

I've done this a few times over at WT

I've done this more times (hundreds) than I can count over the past 10 years of IL2.. Initially I did the testing manually which required several tests to remote pilot errors, than the 3rd partly auto pilot program came out that made life much easier, than when the hacks came out, making use of the zInfomod to extract the data directly from the game made it even easier.. But did require some sanity checking of the values.

 

, and why yes, yes I did prove conclusively some of their planes are absurdly off the mark.

one man's absurdly is another man's acceptable.. I tend to stay away from 'words' that require someone to guess at what they mean and stick to numbers and percentages of error. For example, the ROC 30%+ better than it should be at 15kft, which was (still is) the case for the Bf109K in IL2

 

I intend to perform these tests on both the Bf-109F4 and the LaGG-3 in a few weeks when BoS beta is out for us. It will likely be one of the first things I do since I doubt it will contain much content.

But without the ability to log the data in real time, the best we can hope for now is a screen capture or live video recording of the event for others to review.. Because without that we don't know if the top speed was greater that it should be at such and such altitude due to the FM being implemented incorrectly or due to the pilot being in a shallow dive that he did not realize he was in. In that over the past 10 years of IL2 testing, if I had a dollar for every guy claiming the top speed was in error, only to find out it was pilot error I would be a rich man! Edited October 25, 2013 by ACEOFACES

[BeastyBaiter]

one man's absurdly is another man's acceptable.. I tend to stay away from 'words' that require someone to guess at what they mean and stick to numbers and percentages of error. For example, the ROC 30%+ better than it should be at 15kft, which was (still is) the case for the Bf109K in IL2 

 

I consider a 45m/s climb rate on a plane that should have an 18m/s climb rate a little on the absurd side. I doubt many would disagree. That said, I had many over there argue with me endlessly about its accuracy without ever disputing my test results or my sources. Go figure.

[LLv34_Flanker]

S!

 

 Just make sure to have a flame resistant suit, lots of popcorn(or whatever snack you prefer) and beverages at hand when the release hits the market. Then the massive "FM/DM/put desired issue here" threadnaughts will begin 

[GOAT-ACEOFACES]

I consider a 45m/s climb rate on a plane that should have an 18m/s climb rate a little on the absurd side.

Now that is more like it! Numbers paint a much clearer picture than words like absurd!

 

I doubt many would disagree.

Assuming that is the case! As I noted, without the data being log during the test, one can not say with certainty that the observed error was due to an error in the flight model. As I noted, based on my experience of reviewing IL2 track files, most of the so called flight model errors were actually pilot and or test method errors.

 

Which brings us back to square one of my initial post!

 

That said, I had many over there argue with me endlessly about its accuracy without ever disputing my test results or my sources. Go figure.

Exactly, your going to have nit wits argue with you no mater what.. But, when you do it the way it was done you remove one aspect of their nit wit argument.

 

Just make sure to have a flame resistant suit, lots of popcorn(or whatever snack you prefer) and beverages at hand when the release hits the market. Then the massive "FM/DM/put desired issue here" threadnaughts will begin

Bingo!

 

In summary

 

If there is no way to log the flight data, it makes it very Very VERY hard to say with any certainty that the error is an FM error, in light of the fact that most errors are pilot errors.

 

The only option left is to record (ala IL2 track file or screen capture video) the cockpit gauges and exterior (ala IL2 WW view) during the flight for post test review. Which is a very time consuming process in and of itself.

 

Where as with a log file, once you have the post processing analysis tools (MATLAB, EXCEL, etc) set up, it only takes a few seconds to see if the error is a true FM error or a pilot error.

 

There are good reasons for instrumenting planes in real life during testing, they did it back than and they still do it today. And pilot errors is just ONE of the many reason for doing so.

Edited October 26, 2013 by ACEOFACES

[A_S]

A very interesting page with performance charts and combat reports of that area.

 

http://www.wwiiaircraftperformance.org/

[GOAT-ACEOFACES]

Love that website, have for years!

[wtornado]

It better be as good as HSFX or we will never hear the end of it eh ACEOFACES? hehehe

[GOAT-ACEOFACES]

It better be as good as HSFX or we will never hear the end of it eh ACEOFACES?

One would hope and expect a flight model developed 10+ years later to be as good if not better

 

But, until we have some method of independently verifying the flight models it will be nearly impossible to say with any certainty

 

Which results in some saying 'a plane' is under modeled, some saying 'a plane' is correctly modeled, and some saying 'a plane' is over modeled

 

All within the same thread with no proof provided to support any of the three claims!

 

Heck, even with the ability to log the data (provide proof) there will still be some who make all three claims without providing any proof.

 

But they will be greatly reduced in numbers and easily ignored in place of testing that provided proof.

 

And for anyone who does not belive what I say, simply go to the RoF forms and post a message saying one of the three following things

 

1) The Dr1 is under modeled

2) The Dr1 is correctly modeled

3) The Dr1 is over modeled

 

And than stand back and take note of all the people who show up making claims of FACT for and against while providing no proof what so ever to support their claims.

Edited November 10, 2013 by ACEOFACES

[=IRFC=SmokinHole]

The FM debate infinite.  I am inclined to say it is also pointless.  But that is too extreme.  There are situations where the modeled airplane is so far off reality that the developer is at risk of showing bias unless they straighten it out.  Otherwise they should just press on and ignore the chatter.  RoF was a good example of how silly this discussion could become.  Players were working themselves into a lather debating as fact accounts from novels and very inexperienced pilots.  Outside of a few well-studied models like the D.VII very little real testing was done.  The performance was therefore lost to history and any discussion of it is just hamster spinning.  WWII is a different beast for sure.  Even with all the data out there I still think the developer should be given a long leash.  Just like the pilots of the day, we players must fly the planes we are given, not the planes we think we want.

[GOAT-ACEOFACES]

RoF was a good example of how silly this discussion could become.  Players were working themselves into a lather debating as fact accounts from novels and very inexperienced pilots.

This is true of any flight sim, not just RoF.

 

As noted, even with sims like IL2 that provided means to validate the sim by logging flight data during testing will have those working themselves into a lather and debating pilot accounts to coloring books.. But as noted above, these types can be, and are, ignored if they didn't bother to collect the flight data during the test.

[LG1.Farber]

What happened to the clod curves you made?  My links gone dead

[DD_fruitbat]

If i were a cynic, i would say there is a reason why 777 have never enabled logging of flight data in RoF, and why they wont in BoS...

[GOAT-ACEOFACES]

What happened to the clod curves you made?  My links gone dead

I still have the data I collected during the test, but, I am having an issue with my website.. Seems to be an issue with the National Instruments DLLs I used for graphing the data. I think I may have to switch to MS new graph stuff, which aint bad by the way.

If i were a cynic, i would say there is a reason why 777 have never enabled logging of flight data in RoF, and why they wont in BoS...

I hope your wrong, but, you could be right!