Russians Underpowered, US Overpowered in CMBS?

[A Canadian Cat]

2 minutes ago, IMHO said:

1. As far as I remember it's 122mm. 152 is way too much for 2.4 litres of fuel-air mixture in RPO.
2. One important qualification for the HE - FA comparison. When KBP provides this marketing slogan they use a little trick here. They compare the INDOOR effects of an RPO explosion INSIDE the enclosed shelter vs. HE nearby explosions OUTSIDE the shelter. You can make rough comparison by pulling the Joules of RPO mixture burning vs. the Joules of HE shell TNT contents. And certainly there's no hope for RPO to attain effects similar to 122mm HE if RPO explosion is unconfined. The trick in itself is not that absurd as firing RPO from 150m one can hope to be more accurate than calling a fire support from many kilometres away.

I still have not read @Sophist_13's links but the confined vs unconfined explosion question was one that I had. Thanks for those additional thoughts.

[A Canadian Cat]

10 hours ago, Sgt.Squarehead said:

Have you tried a mobility test on the tanks.....The T-64 came out best in five tests for me, yet it has the worst cross country mobility of the lot. 

That does sound wrong. Tell us about how your test is designed please.

[A Canadian Cat]

10 hours ago, Douglas Mac said:

While I agree that proper testing needs to be done before BF can change things, anecdotal evidence is still important and dont think Sgt.Squarehead should criticized for his methods. I would have thought the test he did would be of value and does demonstrate the US advantage in my eyes. I thinks that what we call a control in science? Sure "real world" testing ie within a battle, is probably better but Sgt.Squareheads method has some value as a control and seems to show US superiority. He gives his method and others can repeat it so why is what he did so insignificant LanL?

Oh yes anecdotal stories are how things get started. Without those you don't know where to investigate next. So to be clear @Sgt.Squarehead's testing is worthless it is just not enough to convince BFC to make changes. He is showing there is a difference and since he feels it feels wrong then it is worth investigating. In order to make changes there have to be a clear demonstration of what the difference is and that it is wrong. Defining if a difference is write or wrong is difficult enough even with good stats from a good test. Without that...

 

10 hours ago, Douglas Mac said:

Anyway thanks Sgt.Squarehead for taking the time to do that and the others who have offered explanations. Much appreciated guys.

Yes, believe it or not putting your foot in these waters is a good thing and is very much appreciated by BFC and volunteer testers.

[A Canadian Cat]

32 minutes ago, IMHO said:

1. Changes in fundamental aspects like spotting capability will have profound effects on the whole gameplay not just tank duels. It will be a different game and that means a lot of retesting so here comes the first point

Yes, but if something is incorrect then that work is worth doing.

 

32 minutes ago, IMHO said:

1. When speaking of the balance it seems to me you somehow assume that the force structures and their application tactics should be the same. Like how many T-90s we need to kill one Abrams. May be RUS and US battlefield tactics are very different so such a comparison is not 100% relevant? May be what should be discussed is how many Msta's shells are required to kill on Abrams because leveraging fire support is exactly the RUS tactics?

Nicely phrased.

[IMHO]

31 minutes ago, IanL said:

I still have not read @Sophist_13's links but the confined vs unconfined explosion question was one that I had. Thanks for those additional thoughts.

Sorry, one more thing I forgot to say, RPO effects are estimated for only one room of limited footage. 50m2 room with high ceilings will make an RPO explosion a "somewhat unenclosed" one.

Edited March 21, 2017 by IMHO

[c3k]

@IanL, thanks for the kind words. (By tagging me, I got a notified about this and have now read through the tempest in a teapot.  )

My thoughts, for whatever they are worth:

1. RPO does seem to be underpowered in the game. (However, the game seems to under power a lot of blast/overpressure effects. The reason for this have a solid foundation (game-play reasons, not real-life), and have been explained elsewhere. The main reason is the game keeping infantry too close together.) I have not tried to run any tests to support whether the RPO should have an increased effect or how to do it, let alone how much it should be tweaked.

2. The tankguntest is a horrible test. Really. Let me delve...

If I want to test spotting between "A" and "B", would you agree that a good test would set A and B 3,200 meters away from targets? "A" would try to spot a golf ball, and "B" would try to spot a barn. No: that is NOT a good test. In fact, all it would test is the time for A to spot a golf ball, and the unrelated (but interesting) time for B to spot a barn. They are NOT trying to spot the same size target.

Similarly, a T90A spotting an M1 is NOT equivalent to an M1 spotting a T90. They are DIFFERENT spotting tests.

Now, let's make the test worse: Let's use a platoon of digitally linked M1's. If M1 number 1 spots a T90, it will share (via the C2 links) that spot with every other M1 in its platoon.

Meanwhile, each T90A spots by itself. Each M1 has the power of spotting by 8 sets of thermals. (And they are better thermals than the T90's.) Each T90A gets one (poorer) set of thermals.

This is a test of how 8 top-of-the-line thermals can spot when compared to a single, distant second place, thermal sight. However, that poor thermal sight gets 10 data points.

Another horrible characteristic of the results posted is how few runs were made. I hate to say it, but it takes about 200 runs to get actionable data.

Another unknown is whether the runs posted (I think I saw 12?), is whether the runs were started from scratch or from a single savegame. If it's a savegame, there may be a "bad seed" which is skewing the results. (The spotting system assigns random start times to every unit's "scan period".)

So, in sum, the test sucks and tells us nothing.

However, in a "save the snowflake's feelings" kind of statement, it IS a starting point to further investigations.   (C'mon, that was meant with humor. If you didn't chuckle, well, maybe it's too close?)

What would I do?

First, find the same target. Pick a BMP or a truck. Whatever. Make that target be the SAME for the T90A and the Abrams. Don't pick a US truck as a target for the T90 and Russian truck for the T90: pick the EXACT SAME unit for each.

Next, only have a single, out of command, normal/normal tank. Load the test battle. Run it, record the data. Now, do it 199 more times.

Next, do the same with the other tank, again out of command and normal/normal.

Now, we have an apples to apples comparison.

(One of the tests I tried to run was to eliminate the extra crew members from tanks to test individual thermals/optics. I created a multi-battalion tank pool and dismounted all the crews. Then I ran them through a sniper alley. I then took all the surviving crews which only had a single member left. I finally got them mounted in their respective tanks. (Try that with 200 tanks. Really. It sucks.) Of course, some crews had the driver, some the loader, some the gunner, etc. I had to sort them so just the gunners were mounted in their tanks. Then I found out that the gunner would take the driver's seat in a mobile tank. I had to redo it and immobilize all the tanks, first. THAT was a fun moment. And, I forget now what I came up with. I just remember the soul-crushing amount of work it took.)

The US Army has invested billions in linking combat units with digital information sharing. It is a huge lethality multiplier. You are seeing some of that in your test. Eliminate it to get a true test of spotting abilities.

TL;DR: I agree RPO should be buffed. I disagree that Russians are otherwise unfairly penalized by design decisions.

Ken

[IMHO]

@c3k,

1. Why 200 tests and not 20 or 2000? I mean you know the variance and distribution rule?  Actually some years ago I ran hundreds of tests of one of the old versions of CMSF engine. I was interested in spotting/accuracy/fire effects for infantry. The variance is actually quite low - the set of outcomes is predefined and quite limited. Certainly many things could have changed since then.
2. I'd guess "bad seed" is irrelevant. If the save predefines the spotting then we will not have different spotting results loading from the same save. And we do. It may somehow skew the distribution depending on the logic but how it does it if it does at all - we don't know. Again from CMSF tests - there's a limited pool of outcomes and the result will be a random selection inside the set. I don't mean that the set is somehow "wired-in" (I don't know, actually) looks like it's the result of the engine algorithms. I have not tested extensively if there's a difference between saves and launching a new battle. I did mostly new battles with very limited saves - back then the pool of outcomes was the same for both cases, distribution - I don't know.
3. Test with 200 tanks is dangerous, IMHO. Since we know that "collective scanning" does improve the results we cannot use single test with 200 tanks as a proxy for 200 tests with single tank. 200 tanks setup will test more of a rule on how units coordinate spotting between themselves rather than single unit spotting abilities.
4. For RUS you dismiss an "ol' and good" radio link. Using linear multiplication for Abrams vs. no multiplication for T-90 you make an implicit assumption that a T-90 crew locates the target then keeps very private about the fact

Just friendly corrections...

PS Actually such a heated argument about spotting may be caused by this limited set of outcomes. Since the spotting is binary - see or no see - it's just some of you may be unlucky more often while others are simply fortunate 

Edited March 21, 2017 by IMHO

[A Canadian Cat]

1 hour ago, c3k said:

@IanL, thanks for the kind words. (By tagging me, I got a notified about this and have now read through the tempest in a teapot.  )

I was counting on that

1 hour ago, c3k said:

2. The tankguntest is a horrible test. Really. Let me delve...

See and you came up with thoughts that I didn't eve have. I knew you would offer better advice.

 

1 hour ago, c3k said:

However, in a "save the snowflake's feelings" kind of statement, it IS a starting point to further investigations.   (C'mon, that was meant with humor. If you didn't chuckle, well, maybe it's too close?)

 

LOL just about lost my coffee. I said I wasn't going to poke more fun but I knew ken would

 

1 hour ago, c3k said:

I then took all the surviving crews which only had a single member left. I finally got them mounted in their respective tanks. (Try that with 200 tanks. Really. It sucks.)

LOL yes it does but not as much as finding a crew's tank under fire. I remember that "remount this battalion of Sherman tank crews while this platoon of PzIVs shoots at you" challenge you posted.  I said to my self - that will be easy I accept that challenge. Yeah, I failed - big.

[A Canadian Cat]

47 minutes ago, IMHO said:

1. Why 200 tests and not 20 or 2000? I mean you know the variance and distribution rule?

I'm not Ken but 20 is low enough that two or three outliers (and there are outliers - by design) could effect the average. 2000 is really a lot of testing and doesn't really change the measured value. 200 is a bit arbitrary in one sense but we have to pick a number that is high enough to avoid unluckily hitting a few outliers messing the average but low enough that humans can do the work.

 

47 minutes ago, IMHO said:

For RUS you dismiss an "ol' and good" radio link. Using linear multiplication for Abrams vs. no multiplication for T-90 you make an implicit assumption that a T-90 crew locates the target then keeps very private about the fact

 

Again not Ken but I don't think he really meant that one T90 will get no help from its platoon mate spotting something I think he just meant that the M1's system is lots faster. I personally no nothing about the specifics of that system but I do know that T90s and everyone else in game *does* get spotting help from other units in command.

 

47 minutes ago, IMHO said:

PS Actually such a heated argument about spotting may be caused by this limited set of outcomes. Since the spotting is binary - see or no see - it's just some of you may be unlucky more often while others are simply fortunate 

That is a big factor. Not only that but we remember when luck tipped against us much more than the other way. That sometimes causes us to feel like its not just luck.

[IMHO]

@IanL, I dare to disagree  

1. There are rules of statistics on how to determine the number of tests.
2. My impression from old CMSF tests were that the engine is mostly logic-driven rather than randomized. That's huge respect to BFC guys - I imagine how much time and effort were invested. But logic driven engines inherently limit the pool of outcomes so you don't need so many experiments to isolate a factor in simple processes. I can tell you there were some cases (albeight basic ones) that one could think are random at first. But in reality you could have EVEN A DISTRIBUTION within ten tests.
3. Again because the pool of outcomes is limited and "streaks of (un)luck" is coomon for randomness it's really quite possible someone can be unlucky for 4-5-6 times in a row and get a bad taste in the mouth 

PS Actually this limited pool was the reason why I decided not to publish the results of the tests. It would have killed the joy of acquiring "the feel" of the game

Edited March 21, 2017 by IMHO

[VladimirTarasov]

Biggest problem with the unbalance is that all US vehicles like Bradleys and Abrams have laser warning systems. If there was an option to for example to make them how they are in standard service it would be great. The American military has superior sensor equipments without the doubt, and they are very organized in what they do, however sometimes the spotting really goes too far. I recall having a T-90 diagonally across a field, behind a tree line in another tree line. And a Javelin team just spots it in a short amount of time as if he was expecting there to be a tank there. 

US forces are great of course, high tier military. But I still believe there needs to be some tweaking to spotting issues. At night the Russian side is at a severe disadvantage obviously, that can stay. But illumination rounds should be implemented that would help alot. Also if artillery damage could be tweaked higher like in real life, that would be great too. This game is still very playable. Developers did a good job.

[Rinaldi]

5 minutes ago, VladimirTarasov said:

...

Wow, welcome back Vladimir. You had us worried with your disappearance.

[A Canadian Cat]

46 minutes ago, IMHO said:

@IanL, I dare to disagree  

No dare needed - go for it.

 

46 minutes ago, IMHO said:

PS Actually this limited pool was the reason why I decided not to publish the results of the tests. It would have killed the joy of acquiring "the feel" of the game

That has not been my experience. The last time I was doing a bunch of tests the first run of ten had what turned out to be three outliers (back luck) in it. If I had stopped at 10 my numbers would have been off. There is plenty of logic and straight physics in the game but there is also subtle randomness mixed in. You could be at least partially right that long strings of outlilers (bad luck) are less likely than in real life but I don't think showing the data would to mess with enjoyment. Of course we could be seeing the same thing and I'm concluding that it is not a big deal but you are bothered by it. That's possible too.

[IMHO]

@IanL, try such a setup: an RPG team, the smallest building in front of it, an OPFOR team/squad inside the building, OPFOR is facing RPG team and is ordered not to return fire. Now RPG team makes one shot at the building, you write down the number of KIAs and WIAs should an RPG hits the building then you relaunch the battle. The idea is basically to isolate damage calculation logic. If I remember correctly the sum of soldiers turned into KIAs and WIAs status is ALWAYS a constant. I run the test something like 50 times though ten were enough

PS

1. Inside the sum of KIAs and WIAs the number of KIAs might be preset as well but I'm not so certain about it - it was six years ago.
2. It makes battles a bit "gamey" since you can more or less accurately estimate the number of OPFOR guys left in the building after multiple RPG hits - due to this constant outliers will be quite rare. Having tested spotting / probability to hit / suppression capability of a team in the building one can predict the damage to one's soldiers should you desire to storm the building.
3. To minimize outliers you do need to carefully isolate the factors. Should you try to estimate the end result of a rush to the building per se - you'd need many tests for sure. But if you isolate then summing up the factors and calculating mathematical expectation is straightforward.

Edited March 21, 2017 by IMHO

[c3k]

IMHO,

Perhaps I was unclear. My T90A vs Abrams 200 test was meant as 200 individual single tank tests, not a single 200 tank test. (My anecdote used about 200 tanks because that was how many I needed to start with to enable the culling of the crews to produce a handful of usable one-gunner crews. Sorry for the confusion.)

As to stats, I'm using the 200 number because that is what is needed for a convincing argument. It eliminates all outliers, has an extraordinary high confidence level, and eliminates any argument about sample sizes. Now, you could argue a smaller number...but that would be a debate and introduce a confidence interval into the results which could be used to undermine your conclusions.

Now, an individual test MAY have more than one tank, and still be valid. It would just have to ensure that no two tanks have any form of communication between them, such that they act, for all purposes, as if they are on the battlefield by themselves.

Ken

[shift8]

There is nothing to see here. The tank with the quantitative and qualitative advantage in optics spots the Russian vehicles first and wins, what a surprise....not. 

The alleged proof of extreme error in the modeling is the assumption that the crew expertise of the T-90AM would magically overcome technological advantage of the Abrams via some kind of magical proportional override, as if crew experience is some kind of mystical force that can offset real-world technical challenges to any extent that seems convenient.

A T-90AM crewed entirely by Chuck Norris clones is about as likely to break even with the M1 in spotting as a M1 crewed by the same clones is likely to start seeing through walls and predicting the future. 

[IMHO]

@c3k,

1. Believe me 200 is no more convincing than 20 or 2000 from a statistical point of view
2. Multiplying ten Gaussian distributions and one say log-normal makes the end result barely different from a simple Gaussian one.
3. If I remember correctly the factors for spotting were: number of eyes, experience / suppression / bla-bla-bla, angle to the target, distance / weather / etc., probability to achieve SOME level of spotting for a given time slot (I guess it works through assigning sectors for a period of time but exact implementation is not so important for predicting the result), probability of "full-spot" vs. "semi-spot". Each individual function was quite simple - the end result, certainly, complex but only if you do not isolate. If you do - it's pretty easy to calculate. Outliers are not due to the complexity of each individual function - they are because the spotting is binary.

Edited March 21, 2017 by IMHO

[Rinaldi]

Suggesting that sample sizes makes no statistical difference is heresy for someone with a background in that form of analysis. It is, to me, a very foolish assertion to make - regardless of how 'rigid' you feel weapons systems act in the game.

Beleive me 20 is far less convincing than 200 or 2000 from a statistical point of view

Edited March 21, 2017 by Rinaldi

[IMHO]

@Rinaldi, well if you make few steps away from textbook Gauss into the probability distribution one has to work with in real life  I had a real life case when 50 mln values were not enough to acceptably approximate the probability function  You believe 200 would have been any more convincing than 20? 

Correction: 3.5mln * 2 * 12=84 mln results and about 1'500 factor values for each result.

Edited March 21, 2017 by IMHO

[Rinaldi]

In the context of CM, yes absolutely. I would not take anything as 'fact' (speculative or otherwise ) with anything less than at least a 100 sample size in game, and that is being very liberal.  It's lead to considerable differences in results from those who did considerably less testing; which basically leads me to throw your reasoning out the window immediately. Presuming I fully understand where you're coming from; you're suggesting that the game really doesn't have many variables and ultimately if x then y in all instances. That would be perfectly sound if I held it as true, but I don't.

 You say the game follows a logic - well yes that's a painfully obvious statement to most. Part of that logic is "standard" deviations (always an odd phrase) that give us something approaching RNG. It's quite observable, and would be ever so much more so observable with an adequate sample size . I wonder if I did the RPG test 200 times in a row precisely as you asked, would the result actually model your hypothesis? Perhaps when I'm home I'll indulge you, one of us will be vindicated. Besides, as Ian indicates below, ordinance effectiveness versus spotting effectiveness are very different beasts; the latter has many more variables (such as the fact that it operates on a cycle) that we cannot control as players.

As I said: I could set my watch to threads of this types. It's not limited to Black Sea. Every 6 months or so there is an inevitable individual with a bone to pick whom swings for the posts viz. the Panther, Russian SMGs (Why are my UBERMENSCH losing to these PPSH armed plebians!?)* or some such other tank or weapon they percieve as under-performing and they give us a very modest test and sample size that doesn't stand up to even a bit more scrutiny. Or even worse, a case in which they are getting their ass handed to them by lady luck in a single instance and decided this was the straw that broke the camel's back.

 

* That thread was almost as entertaining as this one - almost, but not quite.

Edited March 21, 2017 by Rinaldi

[A Canadian Cat]

1 hour ago, IMHO said:

@IanL, try such a setup: an RPG team, the smallest building in front of it, an OPFOR team/squad inside the building, OPFOR is facing RPG team and is ordered not to return fire. Now RPG team makes one shot at the building, you write down the number of KIAs and WIAs should an RPG hits the building then you relaunch the battle. The idea is basically to isolate damage calculation logic. If I remember correctly the sum of soldiers turned into KIAs and WIAs status is ALWAYS a constant. I run the test something like 50 times though ten were enough

OK wow interesting. I have never done testing to determine what casualty rates are like for ordnance X. My tests have been about spotting, chance for track damage running over obstacles, chance to hit a target on the first shot. Those are not deterministic and have plenty of variance. The appropriate sample size for confidence of an average has to vary based on the variability of the event. Clearly if the game only gives you one result ever time you don't need to repeat it. But I can tell you there is plenty in this game that does not repeat the same every time. So a sample size of 20 is not really enough.

<spoiler>

I totally get what you are saying about too much information causes crack in the immersion. Mine is the realization that the icons above the enemy are not centred on the visible soldiers but on the entire team. Which means you can get an idea of the enemy team size and location.

I try to not look too closely.

</spoiler>

[Sgt.Squarehead]

Fair play Ian, sorry if it seemed like I was trying to give you a hard time.....Couple of points, while the Abrams does have a distinct sensor advantage, it does not have an overall visibility advantage.  IIRC the T-72 & T-90 both have superior short-medium range spotting simply because the gunner can participate too. 

My fear is that too much emphasis has been placed on the poor performance of Russian equipment in the two Gulf Wars, to say that it was hardly being handled at it's best in those campaigns would be a monolithic understatement.

If there is meaningful testing that could be done to help straighten this out I'd be glad to do it, my PC is quite capable of running one CM game while I'm editing another.

[A Canadian Cat]

35 minutes ago, Sgt.Squarehead said:

Fair play Ian, sorry if it seemed like I was trying to give you a hard time.....

No worries - thick skin here and all. At least most of the time.

 

Quote

Couple of points, while the Abrams does have a distinct sensor advantage, it does not have an overall visibility advantage.  IIRC the T-72 & T-90 both have superior short-medium range spotting simply because the gunner can participate too. 

I seriously question that. The M1 gunner is not facing backwards or anything. What make you think that the T72 and T90 have better visibility than the M1 in any situation? Serious question. I have no tank experience my "knowledge" is only from what tankers have told me. So far none of them have made the case that a T72 or T90 would have better situational awareness at any range compared to an M1.

 

Quote

My fear is that too much emphasis has been placed on the poor performance of Russian equipment in the two Gulf Wars, to say that it was hardly being handled at it's best in those campaigns would be a monolithic understatement.

What do you base that on? The US tankers that have commented here have repeatedly commented that they recognize that the Iraq crew training was not nearly as good as top tier Russian tankers. I have never seen anything that led me to believe that the technical capabilities of a T72 were under estimated, by BFC, due to performance of a battle group in any war.

 

Quote

If there is meaningful testing that could be done to help straighten this out I'd be glad to do it, my PC is quite capable of running one CM game while I'm editing another.

Sound good. Honestly I have a few closeted concerns about spotting - mostly in close though - not the situations you are talking about. But if you want to do some investigations then I suggest you combine some of what I was talking about earlier: creating multiple test lanes (creating separate lanes - with one tank per lane and use actual terrain of 20-30m high hills to separate the lanes) and what @c3k talked about (having the test vehicles all spotting exactly the same vehicle). If you want to add in having the spotting target firing then don't use trucks some some AFV. Then you can run the test with T90s spotting (say BMP3s) and then M1s spotting the same thing and then do it again but with the T90s spotting firing BMP3s and then M1s spotting firing BMP3s. Then you probably will want to do same at different ranges. I the work goes on and on

Before you get too involved though I suggest you pick one distance and one vehicle to try it out and share the test design first. Trust me that last thing you want to do is create 10 scenarios and run 1000+ runs only to find out some smart ass like me or @c3k (way more likely it will be @c3k cause he is the most experienced at this) poke a whole in the plan and blow up all that work. Been there done that - it makes you . Way better to get feed back on one setup with some small number of iterations - just for discussion - before pulling the trigger on a tone of variations and iterations.

I also recommend you start a new thread for it since this one has a lot of stuff going on and will distract people from what you are up to.

Edited March 21, 2017 by IanL
grammar and clarity

[Sgt.Squarehead]

1 hour ago, IanL said:

I seriously question that. The M1 gunner is not facing back wards or anything. What make you think that the T72 and T90 have better visibility than the M1 in any situation. Serious question. I have no tank experience my "knowledge" is only from what tankers have told me. So far none of them have made the case that a T72 or T90 would have better situational awareness at any range compared to an M1.

It's my recollection that the gunner of the T-72 and T-90 has a 360⁰ periscope whereas the Abrams does not.....This is just a recollection off hand (probably based on something I read at the Tankograd Site), post WWII AFVs are not really my thing (that would be WWII AFVs) so I don't have comprehensive references to check all the details to hand, but I can (& will) check online resources in due course.

8 hours ago, IanL said:

That does sound wrong. Tell us about how your test is designed please.

The test is attached at the bottom of the post Ian, I always try to do this, so we all know what we are talking about.....If you take the time to get to know me you'll find I'm very far from a snowflake.

c3k's comments are enlightening, but not entirely unexpected, his comments regarding the sample size was something I was certainly expecting.  It should be remembered though that most players won't bother to replay a scenario 200 times to see if they get a more plausible result, they'll just abandon the game in frustration and probably slag it off in a forum somewhere. 

As for the Abrams networking, the T-90s are within shouting distance of each other.....I did actually wonder if the sheer density of targets was skewing things in the Abrams favour.  The simple fact is that most neutral observers viewing that test would very probably come to the conclusion that something just isn't right (and I don't think my testing methods would be at the top of their list).

PS - Looks like I may have been wrong about the T-72 having extra periscopes.....It seems it was this quote about the T-80 (and thus presumably also the T-90) that I had in mind:

Quote

Later on, both the T-80B and T-80U turrets placed two TNPO-165 general vision periscopes in  and one TNPO-160 periscope aimed to the right, giving the gunner a good view of his surroundings when needed in addition to helping to improve the lighting condition of his station, which is pretty neat as well.

Keep in mind that in most NATO tanks, the gunner is not provided with any general vision devices at all, but inversely, the station is extremely cramped and amenities are few are far in between. Wider tankers will find it very difficult fitting into the station thanks to the massive GPS (Gunner's Primary Sight), but lankier people will find it decently accommodating, especially since the lack of a turret basket means that he will be able to stretch his legs. If the gunner is short and slim, all the better.

Edited March 21, 2017 by Sgt.Squarehead

[shift8]

If anything the reasons for the failure of soviet gear during the gulf wars has been beaten like a dead horse. The only circles who view the performance of Iraqi armor as representative of soviet armor are people whose interest in it doesn't go beyond the history channel. The horse has been beaten so thoroughly that the grog-land perception is now the opposite extreme, with the assumption always being that every statement positive of American equipment being immediately regarded as either propaganda or uninformed.  

 

Make no mistake, the Russians are very competent and make very good stuff. But the sanctimonious vigor that its defenders have is getting quite old.