Monthly Dev update #2 (May 2022) - Angry Cats Looking for a Dogfight!
Dear Fans,
Dear Followers,
Old & new,
Welcome to this May dev diary for Task Force Admiral.
As promised, we are now on a monthly schedule. This one will be dedicated to air-to-air mechanics.
[h2]Hello to all our new friends! (aka thanks Shermie!)[/h2]
First of all, allow us a shout-out to Sherman on Youtube, who had the kindness to cover us in a video earlier this month. Seems it went quite well (100k+ views, cool!) and there's a lot of you out there who mentioned how they'd love TFA to somewhat bring back the emotions we all had when playing Battlestations more than a decade ago. Well, we can't promise anything when it comes to direct control of units, although it is considered to some extent for future development - but be certain we'll do what we can to have the sort of looks, sound environment & general vibe necessary to look the part!
At any rate, here are the sequences - some of them original, some of them longer versions of scenes you might have seen already. Hope you'll enjoy the ride. As for all our new followers who joined us following this vid, welcome aboard - and let's get to the next item of interest.
[previewyoutube][/previewyoutube]
[h2]Drama in the air: developing air combat and TFA's flying fighting machines[/h2]
We have reached a point in development where we're busy testing - if I might say, even brute stress-testing - existing mechanics so as to ensure that they will work properly, especially in the demanding scope of an actual carrier battle. As you certainly know, these engagements could involve dozens, if not hundreds of aircraft in relatively close distance of each other, all busy with their own mission and priorities. At one point at Midway, literally more than three hundred planes were in flight between Midway, the Kido Butai and the American carrier task forces.
Granted, not everybody was engaged at the same time in the same spot, but still we have to plan for contingencies. In an alternative world, where the US commanders get their act together, Midway's own attack force, say, could have been striking at the Japanese carriers at the same time as all the converging carrier-based raids, while the Japanese were launching their own raids. In such an extreme scenario, what would happen to your computer next?
Well, good news, we ran some tests: seems like your computer will be alright. That's a good start, isn't it! Unlikely 100 v 100 dogfights were set, and our computers are handling that like fairly well. That one thing is a relief already, and it is pleasing to know that this part of the development is on the right track. Engagements of a moderately epic scale (for the genre that is... We ain't UEBS in the air either) will be possible and should allow us to rest easy when simulating some of the most complicated battles of 1942, including on a resource-heavy map like this one which encompasses the Solomons, New Hebrides, New Britain, a bit of New Ireland and a big chunk of Papua New-Guinea.
This also allowed us to explore a bit further the damage system. As you can see above, each aircraft is made of larger main parts (wings, fuselage, tail, engine which all have a certain amount of hitpoint before they finally break down, possibly with catastrophic consequences). Sub-systems, such as the pilot himself, fuel tanks, landing gear, radios, other avionics when present (IFF, navigation homing devices, radars), guns, etc... will be added to each aircraft and can be damaged or destroyed by the occasional bullet or shell. It is, obviously, nowhere close to the sort of detail you would have in a simulation like, say, Il-2, but it is still a good compromise considering we're making a wargame with an air combat element - and not the other way around (we'll mention this again later in this update).
Now, as for how combat proper would look, I guess a video is worth a thousand words. Here are a few minutes captured and put together in a short showcase.
[previewyoutube][/previewyoutube]
So, what you're looking at is our fighter AI interacting on both sides of the aisle, trying to make best use of the aircraft flight models and overall doctrine they're provided with. The pace of the action is deliberate: Task Force Admiral doesn't try or ambitions to be on the arcade side of things. Air combat will be at times murderous, but it's nothing like the sort of overkill you'd see in action movies or more recent action games. We try to inject some humanity in our pixel pilots: they are not omniscient, their actions require a base reaction time, wingmen will try to cover their leaders before going for the cheap kill, etc... We encourage them to exploit the vertical plane and the offensive and defensive maneuvers (split-S, Immelmann, scissors) that make good use of it. That would explain why in this sequence, you see them start the engagement over the clouds and end up on the deck. Besides, when specific conditions are met, complex defensive maneuvers like the Lufberry circle or the Thach Weave, or beam defense maneuver like Thach would call it himself, will be attempted by sections under pressure.
[previewyoutube][/previewyoutube]
Overall, it is not perfect, but it feels real enough so as not to warrant much more work for the time being. We'll fine tune AI behavior based on results once we get an actual battle running from the beginning to the end, including through the calibration and use of varying skill levels for the pilots (Midway-era VF-8, watch out!).
At any rate, for those who did not know about our earlier efforts regarding the implementation of complex tactics, please check out our old video about the Weave - that one ain't nothing new, but if you had missed it, here it is for your enjoyment!
[previewyoutube][/previewyoutube]
In regard of aircraft performance, that one recurring question, we will stick with what is made available to us from the archives. Then again, we don't pretend to hold the holy truth when it comes to aircraft specs - honestly who does? Old Il-2 hands who have been around for a couple decades will know what I am talking about. But we'll try as much as possible to remain faithful to the references we have in our possession.
Our main dev JB has developed a tool that allows him to double-check in-game aircraft performance against actual curves from existing data. Such a feature should also be welcome if you were to mod that thing down the road, whether it is about tweaking the performance of existing aircraft... or pre-visualizing the flight characteristics of new ones.
[h2]History corner: Fighter sections in action, or where game design comes from[/h2]
As small essay might appear necessary, as to explain what you were watching just earlier, exactly. That is, how we proceed when we design our AI behavior, in regard of what references we use, and why we’ll always go with doctrine, reference literature and statistics over a mere “gut-feeling” whenever we can afford to do so.
One thing that might not seem too obvious and lead to some surprise when watching the action depicted in the videos, is the way the said action seems to be based around sections rather than individual planes. Indeed, it so happens that modern takes on World War 2 air combat in our films and video games tend to completely set aside the importance of the section as the actual basic combat element, instead of individual aircraft. Much like naval surface combat is in fact a game of divisions (we'll come back to this in due time), naval air combat in 1942 in the Pacific (I wouldn’t go as far as commenting other theaters) was a matter of tactical sections bouncing tactical sections. Although we could very well indulge in that department and go with the flow, it just did not feel right, especially as a starting point for our game series.
Certainly, what you see is arguably showing near-perfect behavior and discipline on part of our pixel pilots, which was what was expected, yet not always found in actual American pilots, and even less so in their Japanese counterparts. We'll make sure to add some more uncertainty to all of this, especially through our pilot proficiency system. Yet, the section - and formation flight in combat - was not just a fancy impracticable idea born from the desk of officers safely theorizing behind the lines. These combat methods were designed by actual fliers, and flown in combat by then.
In regard of the section concept proper, its point was to precisely enable, well, section combat tactics. These could only be done as long as fighter pilots followed “intensive training in maintaining position in a section formation in every type of section maneuver.” Indeed, as “the section being the basic unit of the fighting squadron, every section must be thoroughly trained to maneuver as a unit at all times.” (USF-74, 2-302 – Basic training)
Section tactics have their use on the offensive, where the wingmen are expected to cover each other during firing passes, but obviously are the condition sine qua non to enable some of the more complex team-based defensive maneuvers, such as the Lufbery Circle or later on the Thach Weave. To be pretty straightforward, you can’t make a circle alone, nor can you weave without another friendly aircraft (or ideally, by the book, another section). These maneuvers were important enough that the Beam Defense Position (as Thach called the Weave) led to the adoption, service-wide, of the divisions of four aircraft made of two sections, instead of the (until then) regular divisions of six aircraft made of three sections, as the earlier model left one section without a role or a cover. Even then, this reform and the tactics which motivated it were only allowed by a prior change in 1941, when the Navy, learning from the lessons of air combat over Europe, had converted the basic fighter tactical unit from the three-plane section to the pair. That’s how important these formations were: they defined the shape of fighter air combat by merely subtracting a plane from a section, or a section from a division. Ordered air combat was no furball – nor was it fantasy, as it was the foundation of the greater success the US Navy would obtain in the air, first in the Solomons with the performance-challenged F4F, then for the rest of the war with its more potent successors.
Agreed, there’s still much to do to give the Japanese counterpart, the Shotai, the same sort of relevant treatment in-game, and take into account the relative yet well-documented hotshot tendency of its pilots. But at the end of the day though Task Force Admiral is - we'll repeat it, sorry - a wargame with a lite air sim component, and not the other around. Beyond the show, it is important for us to focus on the actual results & statistics of the encounters we orchestrate, and put to rest a few popular tropes along the way if need be. In that regard, this leads us to another cliché that might need some clarification too – that is, overall air combat lethality.
From the little that could be collected from the comments made under the videos posted above, some viewers have been worried about how deadly these combats would be if the AI is not able to pick every single opportunity to shoot its guns at the enemy. It just so happens though that the overall death toll of naval fighter air combat over the Pacific in the first half of 1942 is particularly underwhelming compared to your average video game session.
Naval fighter encounters for instance (that is, every time a carrier-based US fighter type met with its Japanese Navy counterpart, be it a carrier or shore-based squadron) between Pearl Harbor and June 1942 can be counted with seven of your fingers alone, actually, as you can see in the table below. Note that all aircraft were not all engaged at the same time or at all, and that bombers were present too (but here only fighter on fighter kills are accounted for). Damaged aircraft are indicated when known in details, and this table does not claim to be overly accurate (it was compiled for the sake of the argument), but gives a good idea of the actual losses. Note that 3 to 4 times as many kills (or more!) were routinely claimed by the airmen on both sides during the very same actions.
For most of the pilots involved in these encounters, it was their very first experience with air combat. Some had not flown in a combat environment at all before these dogfights. In addition, US combat performance during these early days was particularly vulnerable to debilitating gun jams. As these were solved and pilots on both sides learned the ins and outs of the enemy (and the limitations of their own aircraft), and as the elite Tainan Kokutai pilots started to fly over the Solomons at the very edge of their range, naturally, these losses increased. But it is a good reminder that one-sided battles were rare, and definite kills hard to achieve. Most of the excellent pilots involved in these battles never became aces, the luckiest ones had a few victories by July 1942, most none at all – and even though these numbers include bombers and search planes too, bear in mind that most of these kills were not confirmed by post-war research when they involved fighters. Unhinged sky massacres, leading to the obliteration of one side when it came to fighter combat as you know them from your usual multiplayer engagements, simply did not happen back then, that is a fact. And this is what we will base our in-game experience off, as much as we can.
Good sources for these contents are naturally books such as the legendary The First Team duology by John Lundstrom at USNI Books, and Michael Claringbould & Peter Ingman's South Pacific Air War series - and many other volumes over there at Avonmore Books. These are all very good readings which will set the mood just right for our game.
Finally, allow us to wish all Veterans, Servicemen, Servicewomen and those who love them and care for them a warm, restful, kind Memorial Day. We know there are quite a few of you from the US Navy, Marines & Coast Guard - and relatives - among our followers. A few kind words are the least we can do, even though we are not from your place - but still owe you much of our freedom, and have gone through some loss ourselves all too recently. Thank you all for your service, and the service of those you love - and please stay safe.
Take care, all of you.
The Task Force Admiral team.
Dear Followers,
Old & new,
Welcome to this May dev diary for Task Force Admiral.
As promised, we are now on a monthly schedule. This one will be dedicated to air-to-air mechanics.
[h2]Hello to all our new friends! (aka thanks Shermie!)[/h2]
First of all, allow us a shout-out to Sherman on Youtube, who had the kindness to cover us in a video earlier this month. Seems it went quite well (100k+ views, cool!) and there's a lot of you out there who mentioned how they'd love TFA to somewhat bring back the emotions we all had when playing Battlestations more than a decade ago. Well, we can't promise anything when it comes to direct control of units, although it is considered to some extent for future development - but be certain we'll do what we can to have the sort of looks, sound environment & general vibe necessary to look the part!
At any rate, here are the sequences - some of them original, some of them longer versions of scenes you might have seen already. Hope you'll enjoy the ride. As for all our new followers who joined us following this vid, welcome aboard - and let's get to the next item of interest.
[previewyoutube][/previewyoutube]
[h2]Drama in the air: developing air combat and TFA's flying fighting machines[/h2]
We have reached a point in development where we're busy testing - if I might say, even brute stress-testing - existing mechanics so as to ensure that they will work properly, especially in the demanding scope of an actual carrier battle. As you certainly know, these engagements could involve dozens, if not hundreds of aircraft in relatively close distance of each other, all busy with their own mission and priorities. At one point at Midway, literally more than three hundred planes were in flight between Midway, the Kido Butai and the American carrier task forces.
Granted, not everybody was engaged at the same time in the same spot, but still we have to plan for contingencies. In an alternative world, where the US commanders get their act together, Midway's own attack force, say, could have been striking at the Japanese carriers at the same time as all the converging carrier-based raids, while the Japanese were launching their own raids. In such an extreme scenario, what would happen to your computer next?
Well, good news, we ran some tests: seems like your computer will be alright. That's a good start, isn't it! Unlikely 100 v 100 dogfights were set, and our computers are handling that like fairly well. That one thing is a relief already, and it is pleasing to know that this part of the development is on the right track. Engagements of a moderately epic scale (for the genre that is... We ain't UEBS in the air either) will be possible and should allow us to rest easy when simulating some of the most complicated battles of 1942, including on a resource-heavy map like this one which encompasses the Solomons, New Hebrides, New Britain, a bit of New Ireland and a big chunk of Papua New-Guinea.
This also allowed us to explore a bit further the damage system. As you can see above, each aircraft is made of larger main parts (wings, fuselage, tail, engine which all have a certain amount of hitpoint before they finally break down, possibly with catastrophic consequences). Sub-systems, such as the pilot himself, fuel tanks, landing gear, radios, other avionics when present (IFF, navigation homing devices, radars), guns, etc... will be added to each aircraft and can be damaged or destroyed by the occasional bullet or shell. It is, obviously, nowhere close to the sort of detail you would have in a simulation like, say, Il-2, but it is still a good compromise considering we're making a wargame with an air combat element - and not the other way around (we'll mention this again later in this update).
Now, as for how combat proper would look, I guess a video is worth a thousand words. Here are a few minutes captured and put together in a short showcase.
[previewyoutube][/previewyoutube]
So, what you're looking at is our fighter AI interacting on both sides of the aisle, trying to make best use of the aircraft flight models and overall doctrine they're provided with. The pace of the action is deliberate: Task Force Admiral doesn't try or ambitions to be on the arcade side of things. Air combat will be at times murderous, but it's nothing like the sort of overkill you'd see in action movies or more recent action games. We try to inject some humanity in our pixel pilots: they are not omniscient, their actions require a base reaction time, wingmen will try to cover their leaders before going for the cheap kill, etc... We encourage them to exploit the vertical plane and the offensive and defensive maneuvers (split-S, Immelmann, scissors) that make good use of it. That would explain why in this sequence, you see them start the engagement over the clouds and end up on the deck. Besides, when specific conditions are met, complex defensive maneuvers like the Lufberry circle or the Thach Weave, or beam defense maneuver like Thach would call it himself, will be attempted by sections under pressure.
[previewyoutube][/previewyoutube]
Overall, it is not perfect, but it feels real enough so as not to warrant much more work for the time being. We'll fine tune AI behavior based on results once we get an actual battle running from the beginning to the end, including through the calibration and use of varying skill levels for the pilots (Midway-era VF-8, watch out!).
At any rate, for those who did not know about our earlier efforts regarding the implementation of complex tactics, please check out our old video about the Weave - that one ain't nothing new, but if you had missed it, here it is for your enjoyment!
[previewyoutube][/previewyoutube]
In regard of aircraft performance, that one recurring question, we will stick with what is made available to us from the archives. Then again, we don't pretend to hold the holy truth when it comes to aircraft specs - honestly who does? Old Il-2 hands who have been around for a couple decades will know what I am talking about. But we'll try as much as possible to remain faithful to the references we have in our possession.
Our main dev JB has developed a tool that allows him to double-check in-game aircraft performance against actual curves from existing data. Such a feature should also be welcome if you were to mod that thing down the road, whether it is about tweaking the performance of existing aircraft... or pre-visualizing the flight characteristics of new ones.
[h2]History corner: Fighter sections in action, or where game design comes from[/h2]
As small essay might appear necessary, as to explain what you were watching just earlier, exactly. That is, how we proceed when we design our AI behavior, in regard of what references we use, and why we’ll always go with doctrine, reference literature and statistics over a mere “gut-feeling” whenever we can afford to do so.
One thing that might not seem too obvious and lead to some surprise when watching the action depicted in the videos, is the way the said action seems to be based around sections rather than individual planes. Indeed, it so happens that modern takes on World War 2 air combat in our films and video games tend to completely set aside the importance of the section as the actual basic combat element, instead of individual aircraft. Much like naval surface combat is in fact a game of divisions (we'll come back to this in due time), naval air combat in 1942 in the Pacific (I wouldn’t go as far as commenting other theaters) was a matter of tactical sections bouncing tactical sections. Although we could very well indulge in that department and go with the flow, it just did not feel right, especially as a starting point for our game series.
Certainly, what you see is arguably showing near-perfect behavior and discipline on part of our pixel pilots, which was what was expected, yet not always found in actual American pilots, and even less so in their Japanese counterparts. We'll make sure to add some more uncertainty to all of this, especially through our pilot proficiency system. Yet, the section - and formation flight in combat - was not just a fancy impracticable idea born from the desk of officers safely theorizing behind the lines. These combat methods were designed by actual fliers, and flown in combat by then.
In regard of the section concept proper, its point was to precisely enable, well, section combat tactics. These could only be done as long as fighter pilots followed “intensive training in maintaining position in a section formation in every type of section maneuver.” Indeed, as “the section being the basic unit of the fighting squadron, every section must be thoroughly trained to maneuver as a unit at all times.” (USF-74, 2-302 – Basic training)
Section tactics have their use on the offensive, where the wingmen are expected to cover each other during firing passes, but obviously are the condition sine qua non to enable some of the more complex team-based defensive maneuvers, such as the Lufbery Circle or later on the Thach Weave. To be pretty straightforward, you can’t make a circle alone, nor can you weave without another friendly aircraft (or ideally, by the book, another section). These maneuvers were important enough that the Beam Defense Position (as Thach called the Weave) led to the adoption, service-wide, of the divisions of four aircraft made of two sections, instead of the (until then) regular divisions of six aircraft made of three sections, as the earlier model left one section without a role or a cover. Even then, this reform and the tactics which motivated it were only allowed by a prior change in 1941, when the Navy, learning from the lessons of air combat over Europe, had converted the basic fighter tactical unit from the three-plane section to the pair. That’s how important these formations were: they defined the shape of fighter air combat by merely subtracting a plane from a section, or a section from a division. Ordered air combat was no furball – nor was it fantasy, as it was the foundation of the greater success the US Navy would obtain in the air, first in the Solomons with the performance-challenged F4F, then for the rest of the war with its more potent successors.
Agreed, there’s still much to do to give the Japanese counterpart, the Shotai, the same sort of relevant treatment in-game, and take into account the relative yet well-documented hotshot tendency of its pilots. But at the end of the day though Task Force Admiral is - we'll repeat it, sorry - a wargame with a lite air sim component, and not the other around. Beyond the show, it is important for us to focus on the actual results & statistics of the encounters we orchestrate, and put to rest a few popular tropes along the way if need be. In that regard, this leads us to another cliché that might need some clarification too – that is, overall air combat lethality.
From the little that could be collected from the comments made under the videos posted above, some viewers have been worried about how deadly these combats would be if the AI is not able to pick every single opportunity to shoot its guns at the enemy. It just so happens though that the overall death toll of naval fighter air combat over the Pacific in the first half of 1942 is particularly underwhelming compared to your average video game session.
Naval fighter encounters for instance (that is, every time a carrier-based US fighter type met with its Japanese Navy counterpart, be it a carrier or shore-based squadron) between Pearl Harbor and June 1942 can be counted with seven of your fingers alone, actually, as you can see in the table below. Note that all aircraft were not all engaged at the same time or at all, and that bombers were present too (but here only fighter on fighter kills are accounted for). Damaged aircraft are indicated when known in details, and this table does not claim to be overly accurate (it was compiled for the sake of the argument), but gives a good idea of the actual losses. Note that 3 to 4 times as many kills (or more!) were routinely claimed by the airmen on both sides during the very same actions.
For most of the pilots involved in these encounters, it was their very first experience with air combat. Some had not flown in a combat environment at all before these dogfights. In addition, US combat performance during these early days was particularly vulnerable to debilitating gun jams. As these were solved and pilots on both sides learned the ins and outs of the enemy (and the limitations of their own aircraft), and as the elite Tainan Kokutai pilots started to fly over the Solomons at the very edge of their range, naturally, these losses increased. But it is a good reminder that one-sided battles were rare, and definite kills hard to achieve. Most of the excellent pilots involved in these battles never became aces, the luckiest ones had a few victories by July 1942, most none at all – and even though these numbers include bombers and search planes too, bear in mind that most of these kills were not confirmed by post-war research when they involved fighters. Unhinged sky massacres, leading to the obliteration of one side when it came to fighter combat as you know them from your usual multiplayer engagements, simply did not happen back then, that is a fact. And this is what we will base our in-game experience off, as much as we can.
Good sources for these contents are naturally books such as the legendary The First Team duology by John Lundstrom at USNI Books, and Michael Claringbould & Peter Ingman's South Pacific Air War series - and many other volumes over there at Avonmore Books. These are all very good readings which will set the mood just right for our game.
Finally, allow us to wish all Veterans, Servicemen, Servicewomen and those who love them and care for them a warm, restful, kind Memorial Day. We know there are quite a few of you from the US Navy, Marines & Coast Guard - and relatives - among our followers. A few kind words are the least we can do, even though we are not from your place - but still owe you much of our freedom, and have gone through some loss ourselves all too recently. Thank you all for your service, and the service of those you love - and please stay safe.
Take care, all of you.
The Task Force Admiral team.
