Monthly Archives: June 2016

The AR-15 You Should Buy: The Colt 6920

A common question I hear is “What AR-15 should I buy?” This might come from a gun owner new to the platform or someone looking to get more ARs. But fear not! Parvusimperator is here to answer this question for you. For both the new gun owner, and the guy looking to expand his collection, the right answer is the Colt 6920.

Why Colt? Why this old-school looking gun? Why not something cheaper like a DPMS? Why not something with more features people want these days like a ‘mid-length gas system’ or ‘mile-long handguards’?

We’ll get to those. First, a brief bit of background. Don’t worry, it’ll be relevant, I promise. Eugene Stoner developed the AR-15 as a follow on to the AR-10, a lightweight competitor that lost out to what would become the M-141 in the US Army’s competition for a rifle in the new 7.62x51mm caliber in 1957. The AR-15 was designed to use a small-caliber high-velocity round to maximize controllability when firing fully automatic and to be easy to aim across the ranges that infantry combat was generally expected to take place at.2 Anyway, both designs were cooked up while he was working at Armalite, hence ‘AR’, which stood for Armalite Rifle. Armalite was a small company out of Hollywood, California, and it was made to apply the latest high tech aerospace3 materials, like forged aluminum and fiberglass, to the firearms industry. The AR-15 was developed for a contract for the US Air Force’s airbase security forces, but Armalite didn’t have much in the way of manufacturing capability on their own. So they sold the design and the rights to it to Colt, who had plenty of arms manufacturing capability.

This means Colt has what’s called the Technical Data Package, or TDP. It’s the plans for the gun in the sort of nauseating detail that only engineers can appreciate. Materials list, dimensions, and crucially, tolerances. Colt has been forced to send this TDP out to other manufacturers for the process of making M4s for the US Military, but those other manufacturers are contractually forbidden from using the TDP to make guns for civilians. So only with Colt are you getting everything as you’re supposed to be. This is most important with little things involving tolerances. Everyone else can reverse engineer the dimensions, but tolerances are harder to come up with if you don’t have the plans. Which means Colt rifles are going to have a correctly sized gas port,4 among other things.

Colt makes M4s for the Army, and given the desire to not fuck up this crucial contract, they’re going to make them right. So they’re going to make your rifle right too, because it’s not worth the trouble to set up a whole new production line. What does this mean for you? Well, it means that all the critical fasteners, specifically the castle nut and the screws for the gas key are going to be properly staked. Staking is deforming a little bit of metal to prevent something from backing out. You could “just use loctite” on the castle nut, I suppose, though those threads are a trifle fine and not really suited to it. Loctite won’t work on the gas key screws though, since that’s where a ton of heat is going. And if those screws come undone, you’re gun is junk. And fine, on the range, that’s just annoying. But it could mean you’ve lost a match. Or, if you’re using the gun professionally, you could end up deader than Elvis. Don’t end up a dead loser. Stake your gas key screws. Which Colt does for you.

You’re also paying for that US Military grade QC with Colt. Which means a Colt rifle is less likely to have out of the box issues. They build their carbines right and inspect every one.

All that said, there are some downsides. The 6920 has exactly two differences from an issue M4. First, it’s got a 16″ barrel to comply with the damnable NFA laws about barrel length. Second, there’s no burst fire/autofire functionality. Those are the only differences. This means the stock is pretty basic, the pistol grip is that infernal A2 jobber that I hate, and you get boring round handguards. Fortunately, the stock and pistol grip are easy to swap out, and that’s fine, because those are intensely personal choices. The handguard contains two heat shields, as is right and proper and standard issue. But it predates all those cool accessories, and it’s not free float5. You also get a fixed front sight block. On the one hand, that’s good, because it’s held in place with two pins. It isn’t going anywhere. On the other hand, that’s going to get in the way of all kinds of handguards you might like to mount.

So even though I’d caution that the stock handguards handle heat of a good extended shooting session very well, and you probably don’t need to attach a ton of crap to your carbine, and that the free float obsession is silly, and that the money spent on a fancy handguard might be better spent on ammo or a class or a good optic, you’re still gonna want one. No problem, I’ll enable you and tell you how to make it work. You’re almost certainly going to need to remove the stock barrel nut, or at least the delta ring and handguard cap, which is going to entail removing the FSB among other things. No problem. When you’ve removed the FSB, you can actually cut the sighting portion off with a hacksaw, file/dremel it to fit under your new handguard without spoiling the gas-collecting bit, and then refinish it with some grill paint. It’s so easy a caveman can do it. Seriously, it’s super easy. Don’t worry if it doesn’t look professional, it’s going to be under a handguard anyway. Make sure it fits, don’t cut the gas part open, and you’re fine.

Colt even will oblige you by doing this from the factory on the 6920 OEM2 model. This rifle ships with no stock, no handguard, no delta ring, no handguard cap, and no trigger guard, so you can swap these parts out yourself. The gas block is cut down and reparkerized from the factory. (This is a nicer way to do it than using grill paint, but grill paint is way easier and cheaper for the amateur). The OEM2 model still has the A2 handguard, because that holds in the spring and detent for the safety, so it has to ship with something there.

As for midlength gas, I don’t get the fuss. The rifle and carbine length gas systems are proven with an absolutely ridiculous number of rounds. It will work. Always. Midlength seems to work too. It’s not as proven. It might be “softer shooting”, but that’s very subjective. Plus, we’re talking about 5.56mm here. This is not an elephant gun. This is an easy rifle to shoot. Really. It is. And just in case, you can always add a comp to reduce muzzle rise that little bit.

So, what have we learned today, class? Stop overthinking it. Stop worrying about it. Just buy a Colt. And ammo. And quality mags. And get some range time.

1.) It’s still probably the best battle rifle of the era on a technical basis, but that’s another story.
2.) For more on small caliber high velocity rounds and why they’re great, see here.
3.) Armalite was funded by Fairchild Aircraft. Back in those days, California manufactured actual things like airplanes, and not just stupid.
4.) Colt does not drill out a giant gas port so that your gun will function correctly with crap ammo. The US army doesn’t use the worst reman and lowest quality dumping ground ammo. Maybe you shouldn’t either.
5.) Insert loud sigh here. I don’t understand the obsession with free floating for 95% of rifle shooters. Most people shoot at a range of 100 yards or less, with at best milsurp M855 or M193 ammo, or possibly that underpowered crap I mentioned above, at targets rather larger than a man’s torso, but god forbid their handguards touch the barrel. Seriously, it doesn’t matter. There are more important things to worry about and bigger fish to fry, like proper technique and practice. And if you’re sitting on a bench trying to split a gnat’s ass at 100 yards with your entry-level carbine and cheap chinese knockoff optic, you’re also being dumb. Get off your butt.

Meet Maryanne, a Product Improved SCAR 16S

So I’ve spent some quality time shooting my SCAR 16S, which has been named Maryanne.1 And the more I shoot it, the more I like it. There are a lot of good things here, and some that I’ve changed, because I can’t not tinker with something. So let’s go through what I’ve found, what I’ve changed, what I’ve kept, and what might change in the future.

The more I shoot Maryanne, the more I like that charging handle. No, it hasn’t hit my thumb yet. Still not sure how that keeps happening to other people.2 Anyway, as I mentioned earlier, locking the bolt back for a right handed shooter is super easy. Grabbing the charging handle, palm-down, and pulling the charging handle back puts your support hand thumb in perfect position to hit the bolt catch. Easy. This makes administrative handling and malfunction clearing3 a breeze. There are a bunch of aftermarket charging handles out there, but I don’t see myself buying one. I haven’t scraped my knuckles on anything yet.

Also, the reciprocating charging handle makes it easy to tell if the gun has stopped firing because of a malfunction or because the mag is empty, without having to rotate it for a chamber check. And, if you like a forward assist, it’ll do that job too. Reciprocating charging handle can stay.

I also like the safety, which is a 45-degree jobber, as opposed to the 90 degree design of a regular AR. I like it better because it’s a bit faster, and indulges some laziness. There are aftermarket safety levers available, but I haven’t bothered with any of those yet, because I can’t tell how they improve upon the factory part. What am I getting with my money besides a tiny bit that says “MAGPUL” on it?

I haven’t touched the factory stock either. Again, I don’t see the point, the factory unit works great as it is. It even has an adjustable riser. And it sits against my shoulder, and it’s adjustable for length of pull, and it folds. Great! 5.56 doesn’t need much of a recoil pad. Also, alternatives tend to be even uglier, so why bother? Maybe if someone made a shoulder thing that goes up for it, I’d get one so I can have another feature to make the antigunners pee their panties, but no one has made one yet. I’ll have to settle for collapsible and folding, as opposed to collapsible, folding and “goes up”. Oh darn.

You’re probably wondering what the devil I have changed. Well, the pistol grip for starters. I hate the A2-style pistol grip. It’s too small and has a nub in a stupid place. Stupid cheap thing. I’ve replaced it with a TangoDown BG-17 grip. This grip is curved and large, for those of us with big hands. It fits my hand great and is comfortable. Plus it puts the distance to trigger in a great spot. It’s a more traditional grip angle though. I don’t understand the ‘vertical style’ that a lot of newer ARs have, but maybe that’s because I’m not kicking doors with an SBR. Shrug. The important thing is that I have a grip I like.

I also replaced the stock gritty trigger with Geissele’s fancy Super SCAR trigger. Trigger replacement on the SCAR is a bit of a pain, but I got there eventually. The end result is a lighter, crisper pull. It’s not too light, and it makes shooting the gun more fun. I’m a civilian. I get to enjoy luxury triggers. This thing is now totally awesome. It feels a lot like my other Geissele triggers, which is fine by me. Mr. Geissele does good work.

I’ve played around with rail covers and the like on Maryanne. I’m currently running Magpul XTMs on the side rails and a Magpul AFG2 on the bottom rail. I’m not sure about this setup. I haven’t shot the rifle to get a feel for how much I like the AFG2 on the handguard, and the XTMs are a pain to work with. I might replace them with the Tango Down rail covers or something. I don’t know yet. It’s easy enough to mess with.

I have not gotten an extended rail yet, because I don’t see the point. Maybe it’s because I don’t have ape-arms, but I haven’t figured out why people like to grip the gun super far out on the handguard yet. Plus, there’s plenty of room for a light if I need to mount one. No extension means I don’t spend money on that and the weight doesn’t go up, woohoo.

The iron sights on Maryanne are really well done. I’m not a big iron sight guy, but these are as good as the high-end KAC sights you might put on your AR if you’re into fancy builds. Range adjustable and everything. Of course, I went optic, because duh. Right now, I have an Aimpoint Comp M4S on there in a Larue QD mount. The Comp M4S is Aimpoint’s flagship red dot. It’s got a 30mm tube, is basically unkillable, takes AA batteries, and has a battery life of about a hojillion hours on a medium setting. I don’t turn it off, and change the battery every few years, if I have to, I guess. It’s easy to adjust brightness, super quick like all red dots, and works great for the generally short ranges I find myself shooting at. Especially in wintertime, I’m stuck on short indoor ranges. Plus, I really don’t like sitting on a bench like a loser and trying to get my groups to be the size of a dime at 50 yards. I have better things to do than carbine masturbation, thanks.

Red dots are simple and cool. We’ll see how it shakes out at some multigun or carbine classes in the near future. I might also get fancy and get an Elcan SpecterDR 1x/4x for it, get some magnification for target identification and those longer shots. I love the 1.5x/6x that I got for another gun project, and the 1x/4x is a little lighter and a little cheaper. Also faster on the low end, because actual 1x. The SpecterDRs some prisms to switch magnification from low to high super fast, because intermediate magnification levels are barely used on most low power variable optics. So, we’ll see. The SpecterDR is a bit heavy, but the mount is built in, so it’s weight is pretty competitive with equally-tough conventional low-power variable optics. I will keep you posted on how the red dot does and if I do go Elcan.4

1.) Yes, I know Fabrique Nationale de Herstal is a Belgian company. And yes, final assembly was completed in Fredricksburg, VA because US firearms import laws are stupid. I do not care. Her name is Maryanne. Feel free to argue with the business end.
2.) If you know how this is happening, please mail a picture of your support hand grip technique to:
c/o I’m a dumb idiot who can’t be arsed to run my gun properly
Ste. 213
82 Stupid Moron Drive
Sewickey, PA 15143

3.) Maryanne hasn’t malfed on me yet, but I’ve done some drills to learn the full manual of arms.
4.) Quick, possibly helpful note on Elcan. Elcan is short for Ernst Leitz Canada. You might Ernst Leitz GmbH as Leica, super awesome camera company. So Elcan makes some phenomenal glass, even though they’re not actually in Europe. They also make the SpecterOS, which is like an ACOG with way better glass. They’re what happened when the Leica guys didn’t want to work for the Soviets after the Second World War. Unlike most optics companies that you’ve heard, they’re a military contractor first and foremost, and are a subsidiary of Raytheon. They’re not known for super crazy long warranties or fabulous customer service. Sorry. Not that you’ll be able to kill a Specter without explosives or shooting it.


The sun has risen on a new day for my friends in the UK. I would like to congratulate them on choosing independence, control, and self-determination over pie-in-the-sky notions and an unseemly, obnoxious bureaucracy.

It is good to see democracy triumph. Congratulations!

Fishbreath Watches: Zootopia

I went into Zootopia with moderate expectations—I had heard good things, but even as a fan of animated movies, I wasn’t expecting anything to boggle the mind. Ultimately, I was still expecting a kids movie with crossover appeal.

This is not what Disney delivered. Zootopia starts out a cheerful-seeming puff piece; by the end, it has taken on an almost noirish feel, telling a pitch-perfect buddy cop story in an exquisitely-detailed world. I can’t say much more without giving away plot points, and the plot is so well-crafted I would feel guilty doing so, but really. If you’re skeptical, fear not: it is not what the previews suggest it is. It’s much better. Go see it. I have not yet heard anyone credible claim to be disappointed.

Done? Cool. I’m going to employ a piece of sorcery most arcane known as ‘the fold’ to hide possible spoilers from casual readers. (Very minor—I’m extremely allergic to spoilers. If you don’t mind non-specific remarks about the flow of the plot through the three acts, you’ll be fine.) Join me on the other side.

Continue reading

OpenTafl: micro-optimization!

Tonight, OpenTafl saw its first micro-optimization!

Let me start from the beginning, though. OpenTafl has a static Coord type, which holds precalculated information about coordinates: what their index is, where they are on the board, and so on. Before network mode, OpenTafl updated the Coord type at the start of each game to correspond to the board size for the selected rules variation.

Astute readers may have noticed the problem. On the server, multiple games may be ongoing, on boards of different sizes. The Coord object should therefore have all the information required for any size of board. The network mode patches included a rewrite of Coord to hold all the information at once, indexed in maps by board size. This solved the problem, but yielded about a performance hit of about ten to fifteen percent. (Obviously, the Coord code paths come up a lot.)

At first, I wasn’t going to worry about it. Then, I had a little downtime this afternoon and decided, “Well, I’m not going to get anything else done right now.” Firing up the old profiler, I found a hotspot to work on.

When exploring the game tree, OpenTafl caches information about how pieces can move. (Calculating it out every time is computationally expensive, or at least more computationally expensive than can be justified doing it every time.) Clearing this cache involves zeroing an array. The usual way to do this is to loop over the array and set everything to zero. (Whether you choose a library function or write it by hand, this is usually what happens.) Unfortunately, that ‘reset’ function, according to the profiler, took longer on aggregate than any other single operation, including such expensive calculations as shieldwall detection and surrounding victories. This would not do.

The solution? Well, although Java’s library methods for array filling do not use native code memory copying, Java’s library method for copying one array to another does. The titular micro-optimization, therefore, was to set the first slot of the array to the desired value, then copy that to the second, then copy those two to the second two, then copy those four to the next four, and so on until the array is fully filled. Since this works directly on system memory, it turns out to be faster.

How much faster? To answer that question, I wrote a little benchmark function for OpenTafl, which runs a 30-second brandub search, a 30-second tablut 9×9 search, a 60-second Copenhagen search, and a 60-second Tablut 15×15 search, which verified the 10-15% slowdown from pre-network to post-network. Fortunately, the move cache optimization yielded a speedup of about 15%, bringing us right back to where we were before. Et voila.

WS-2: Heavy Rocket Artillery

The Russians have done a good job reminding everyone (or at least, everyone watching/paying attention/sending observers like good military intelligence types) that rocket artillery is awesome, and those “icky” cluster warheads are super effective. Even if the hippies hate them.

But of course, if you are worried about the Russians, say, or anyone else, you might want some rocket artillery of your own. And you might want to buy from the Russians, but they might be whom you are worried about, and supplies might be problematic in the event of war. And BM-30 is dreadfully expensive. You might try buying American, but they only make HIMARS these days, which is half an M270. More problematically, they won’t sell DPICM cluster munitions to any new customers, and it’s questionable if they’ll keep selling the stuff to old customers. You could buy guided rockets with unitary warheads, but they are dreadfully expensive. And then you’d want small caliber rockets like the old BM-13, so you could get more than six on a truck.

Ugh. Terrible options. Fortunately, artillery rockets aren’t all that sophisticated weapons, so we could source them elsewhere. Throw in a bit of licensed production and we’d be all set. There are several countries that make their own rocket artillery setups, but today we’re going to go to China.

Ordinarily, we might be a little wary of dealing with China in such matters, but since we’re buying, it’s not like they’ll be able to copy our stuff. Plus, they might be very threatening to Japan and South Korea, but there’s a convenient Russia between us and them. Quite the buffer. Let us see what they have, shall we?

Enter the WS-2. The largest multiple rocket artillery system in the world. It might even be classified as a short range ballistic missile. But shut up, you can get a truck-launcher that holds six missiles, so it’s totally an MRL. They’re just really, really big rockets. They’re 40 cm in diameter and 7.3 meters long. The standard version has an inertial guidance system and a range of 200 km, and a version with GPS/inertial guidance and 350 km range is also available. Perfect for bombarding Taiwan from across the Strait of Formosa. Or pounding enemy concentrations.

What warhead types are available on these monsters? It carries a 200 kg payload, which is quite a lot of high explosive. And HE/Frag and HE/Frag/incendiary unitary warheads are available. But CEP is 600 meters, and that’s not very good unless you’re trying to level Taipei. Or Grozny. So let’s talk things that will make hippies cry.

Warhead option one is the small bomblets. Sino-DPICM, if you will, though these are a bit larger. Each bomblet is a HEAT/Frag munition, rated to penetrate 85mm of RHA steel armor and having a lethal frag radius of seven meters. 540 bomblets are carried in each rocket.

Option two is the bigger, anti-armor bomblets. They still have a fragmentation shell, but add an incendiary component good for a four meter radius. The shaped charge is much bigger, being rated for 180mm of RHA steel armor. I’m not sure if these are guided or not, but it’s easy enough to modify these to work like SADARM/Bofors Bonus/etc. 61 are carried per rocket.

Option three is a unitary thermobaric warhead. It’s good for 29 psi of overpressure at a distance of 25 m from the warhead detonation.

So it’s a big, cool system. Note the obvious shortcoming of a rather large minimum range of 70 km. Also, there are no mine dispensing rockets as of yet, but I’m sure we can work something out. Probably oughtn’t be the only rocket artillery, but it’s a nice oversized, long-range shotgun for the battlefield commander all the same. We’ll probably mount the launcher on a truck chassis that’s already in our inventory rather than add a new one for this.

Resurrected Weapons: Sprint ABM

Suppose you were to take on the challenge of defending against ballistic missiles. The big ones, mind–intercontinental ones with thermonuclear warheads. You might conclude that you’d need a layered defense, with different missiles to attack the ICBM in flight. Of course, the trickiest interception problem is the goaltender’s, i.e. the last line of defense. How to intercept a missile when your shot is the last one is a really tricky problem, and one such solution is the topic of today’s resurrected weapon post.

Behold, the Sprint antiballistic missile system!

It was a relatively short ranged system, with a claimed operational radius of 40 km, and a flight ceiling of 30 km. But that’s to be expected. It’s supposed to be the last attempt to stop an incoming warhead. Of course, since other missiles were to have had the first go, the incoming warhead was beginning its decent. In order to stop it, Sprint had to be mind-bogglingly fast.

The Sprint missile was a two-stage affair, which accelerated at 100 Gs. This sustained acceleration would turn any human passengers into paste. Good thing it doesn’t have any. Sprint would go from zero to Mach 10, or 3.4 kilometers per second, in under five seconds. Given this tremendous velocity, it would intercept a target it’s maximum interception altitude of 30 kilometers in less than fifteen seconds.

In order to make this speed work, Sprint had a number of interesting features for the time. It was cone shaped, and was sheathed in an ablative coating to withstand the extreme1 temperatures generated by the missile. The silo doors were blown off by explosive charges, and the missile was kicked from its silo by a second set of explosive charges, pushing a large piston. Once clear of the silo, the first stage burned for only 1.2 seconds before dropping away.

Sprint had a novel and rather dirty way of defeating incoming warheads. Sprint’s payload was a W66 enhanced-neutron warhead. This low-yield warhead was designed to spray high-energy neutrons, to disable the electronic systems of the incoming warhead, or to cause the warhead to fizzle2 prematurely. Of course, this neutron blast, and the resulting possible fizzle, is not exactly clean, so Sprint was intended for use as a terminal defense system for ICBM silos.

Guidance of Sprint was also a difficult challenge. To keep fragile electronics out of the missile, a large and powerful radar set was emplaced on the ground, and a radio command guidance system was used. At the speeds Sprint traveled, it would be enveloped in a plasma sheath, which would make radio communication difficult. To get around this, the radio beam was made very narrow and very powerful.

So what do we think? As is, it’s kind of specialized. But there’s a market for ABM systems these days, and it bears some further testing with a more conventional fragmentation warhead, and possibly more capacity for maneuvering at speed.

Verdict: Project approved for further research funding by Borgundy Air Force Procurement Board

1.) About 3,400 degrees Celsius or so. Very hot. That speed comes at a price.
2.) A technical term in this case for a subcritical nuclear reaction.

OpenTafl v0.3.1.0b: spectator mode

OpenTafl hit another major development milestone over the weekend: spectator mode! I believe that makes OpenTafl the first real-time-viewable location for online tafl play in the world.

I mentioned in one of my last tafls post that OpenTafl’s architecture has made many of these features extremely easy. Before I let you go today, I wanted to bust out the old object relation chart builder thingy and talk about what I mean.

This is the rough structure of OpenTafl, as relates to the actual play of games. The Game object (along with its children, not shown here) encapsulates the rules. It has methods to play the game: you can ask the Game object who is up to move, and give it a move (in the form start space, end space) to advance the state of the game. It also handles the game clock and a few other bookkeeping tasks.

The CommandEngine object wraps the Game object, and with objects of type Player, handles all the upper-level bookkeeping: sending events to the UI, tracking which player corresponds with the physical human player, and so on. Higher-level components host a CommandEngine object, and correspond to one or more of the Player objects.

When the CommandEngine wishes to notify a higher-level component that something has happened, it uses the Player object. The higher-level components interact with the CommandEngine primarily through the Player object, sending moves and receiving their results through that interface.

Let’s look at an example: a human player, playing a network game against an AI player running on a different computer.

Start with the human player. He’s sitting at the keyboard, looking at the in-game UI. He makes a move. The GameScreen on the screen has a CommandEngine. One of the Player objects is a LocalHumanPlayer, which the GameScreen uses to deliver the human’s moves to the CommandEngine. The CommandEngine delivers the move to the Game and gets the result. The move is good! Hooray! The CommandEngine takes the result and delivers it to the other Player’s opponentMoved method.

That Player is a NetworkClientPlayer. It has a reference to the ClientServerConnection which corresponds to the server, and its opponentMoved method is wired to ClientServerConnection’s sendMove method. The move goes out into the ether.

At the server, a ServerClient receives the packet. The ServerClient has a NetworkServerPlayer as its Player.The ServerClient sees that it’s a move packet, so it parses the move and calls the NetworkServerPlayer’s onMoveDecided method, which sends the move to the CommandEngine associated with the game on the server to which the client belongs. That CommandEngine sends the move to the Game, then gets the NetworkServerPlayer belonging to the other client and calls its opponentMove method. That NetworkServerPlayer sends the move to its ServerClient, which creates a move packet and sends it.

Now, remember that this client is an AI. Does that change anything? Not really! The move packet arrives at the client’s ServerClientConnection, where it is parsed and sent to the connection’s NetworkClientPlayer’s onMoveDecided method. From there it goes into the CommandEngine. The AI plays through a LocalAiPlayer, so the move result goes from the CommandEngine through the LocalAiPlayer’s opponentMove method. Finally, the CommandEngine tells the LocalAiPlayer that it is expected to move, and the AI begins its work.

You’ve no doubt spotted how easy this makes extra functionality. The CommandEngine doesn’t care how the player ultimately makes its moves; it only cares that the player eventually responds to waitForMove by calling onMoveDecided. As such, going from simple human-on-human play all the way to networked play with spectators has required zero major architectural changes. As I said last time, robust.

You may have noticed that I wrote about AI playing over the network. This is not yet implemented, but it is the remaining headline feature for v0.3.x. I hope to build a proper, headless AI-only client, which can be started and run in the background. Ideally, it will have these features: first, it can log in and create games, with or without a password. Second, it can be passed an opponent name on the command line to join and play automatically. Third, after playing, if it created a game originally, it should leave and create a new game. Fourth, it should optionally save records of all of its games on the machine running it, so curious AI authors can investigate how it played. The above behaviors should be controlled by command line switches.

Writing it out, I don’t foresee any of those features causing too much trouble. It should be available soon, and when it is, you can expect to see a few OpenTafl AI players on the intersect server at any given time.

Parvusimperator reviews the T-15 HIFV

Fishbreath wrote an excellent bit defending the BMP-3, the traditional, rather lightly-armored IFV. You should go read it here. In it, he’s replying to my discussions in favor of heavy APCs like Namer, and my own design sketch for a heavy IFV. But I can hardly leave it there. I should reply to him, because argument is interesting and fun. Plus, recent developments bear some pondering.

As you are no doubt aware, the BMP-3 is Russian. Duh. It’s from a similar school of thought as the previous BMPs, heavily armed, cramped, lightly armored. It’ll float. It’s easy to move. Splendid.

Or is it?

I’ve mentioned before the Russian experience in Grozny, but it bears repeating. It was a bloodbath. BMPs were deathtraps. Poorly trained conscripts sent in died in droves because Chechen fighters had thought carefully about urban warfare tactics. Fuel is stored in the egress doors on the BMP-1 and BMP-2. This fuel is supposed to be the first used, but that didn’t always happen. So if the rear was hit, fuel would get sprayed all over the dismounts. And it was likely on fire. Not fun. Plus, given that there’s so much ammo in the BMP, basically any penetration of the turret armor or the forward section meant that the ammo gets hit by the shaped charge jet and blows up too. So, soldiers took to riding on the outside.1

This defeats the purpose of an enclosed vehicle. If they’re just going to ride on top, why not have something like the Sd.Kfz. 251, which had no roof. Of course, that leaves infantry vulnerable to machine gun fire and artillery fragments. Plus, that cold. I hear Russia has a miserable winter. They could suck it up and enjoy their superior deployability. A BMP-3 can float, you know.

This brings us to the T-15. It’s built on that same Armata combat platform as the T-14, except the engine is at the front now, where it belongs for vehicles that carry troops. Two things are of note. First, it’s got an unmanned turret, second, it weighs about as much as the T-14. That’s tank-level protection right there. The Russians have agreed with the Israelis–if you have a vehicle that’s going to get shot at like a tank, it should take hits like a tank. Even outside of urban areas–the Donbass is hardly suburbia. Like most modern armies, the Russians have become much more casualty sensitive2, and shooting an ATGM–even an old one–at a BMP is a great way to inflict casualties.

Armor on the T-15 is typical Russian–a reasonably thick steel and composite structure under a lot of ERA. The reactive armor is a new type, of course, but it’s not clear how good it actually is, because no one has had a chance to shoot it yet. Still, it should be good, since they have lots of experience, and overall protection should be on the order of the protection level of the T-14 given the weight, layout, and the remote turret which I’ll discuss in a moment. Further protection is provided by a hard-kill active protection system, the Afghanit. This system is also in use on the T-14, though again it hasn’t been tested. Not being a fluent reader of Russian, and with the Russians generally keeping things quiet, I don’t know how it compares to other Western competitive systems. I’d guess it’d be similar in performance to Trophy as far as reaction times go, but that’s speculation. They do have a reasonable number of tubes per side, unlike so many western designers who think two per side is enough (it isn’t).

On to the turret. The T-15 has an unmanned turret, just like the T-14. It’s all contained above the turret ring, because a traditional turret basket removes space for dismounts, and carrying dismounts is the T-15’s primary mission. The T-15’s turret is well thought out, and I’m a big fan. It’s got a 30mm cannon with 500 rounds split between a 160 round box and a 340 round box to accommodate the double ammunition feeds. The autocannon is capable of high angle fire, perfect for hitting top-floor rocket teams. The coax gun is in the usual 7.62x54R caliber, and there are 2,000 rounds available. Four tubes for Kornet-EM ATGMs are provided, two on either side of the turret. Kornet-EM has a tandem-HEAT warhead with further improvements, plus automatic command line of sight (ACLOS) guidance, which is an improvement over the usual SACLOS. The gunner has thermal and day sights, plus a laser rangefinder. The commander has his own independent sight which appears to be a duplicate of the gunner’s.

The turret does not appear to be well armored. Since support fires appear to be a secondary tasking, the lack of protection and ease of knocking out the turret is probably not a major issue. I’m fine with this overall, for weight and cost reasons. The IFV’s weapons are much less critical than those of the MBT. For similar reasons, I’m more willing to accept an unmanned turret on an IFV, since any loss of effectiveness is to a secondary mission. Plus, it allows for more hull protection and a full load of dismounts.

The 100mm gun is gone from the T-15. It’s not really needed, as the T-15 has proper, modern ATGMs, and the extravagance is just going to take up turret room which is better spent on more 30mm.

The T-15 carries nine dismounts, plus a crew of three. I do not know enough about Russian tables of organization to know if this is a full squad, but it’s reasonable. It would hold about any current western squad I can think of. This is very good, and is a lot easier than trying to split squads across vehicles.

So what do we think of the T-15? I love it! No, really. Here’s a vehicle that’s on my side in an argument with Fishbreath, and it’s even Russian! They’re replacing their BMP-3s with something much more to my liking. Further, its capabilities aren’t really available anywhere else. We might be able to get close with some modifications to the Namer, but that’s a project for another time. Out of the factory, this is the only HIFV game in town. And I’m a huge fan of the concept, even if I might prefer some minor tweaks. I would prefer missiles with a top-attack profile, and possibly some alternative sensors depending on the particulars. But those are relatively minor points.


1.) Russian and Ukranian troops are doing this in the Donbas now too. BMPs are still deathtraps. They were deathtraps in Afghanistan, deathtraps in Grozny, and they’re deathtraps in the Donbass.
2.) Though the Russian Army is still a conscript one.