To State the Obvious

[vettim89]

For those that have a beef with AE, here is the answer: Write your own Mod,

That's how it has been down since the WiTP days (perhaps all the way back to UV?). With WiTP there were several groups working on a variety of projects. Some of these projects went through multiple evolutions - I think RHS was on version 7 (?). We had CHS, RHS, Big B, EA, and many others. Some of these were the work of a few or even one person but others were a team approach.

Point being if you don't like what you see, get your minions/friends together and get to work. If you think there are parts of the system that don't work or are not historical then mod away. That has been the solution from day one. Yes there are some code changes that you may think need to be addressed but the heart and sole of this game is the database. In fact, there has been more than one clever person who has tweaked the database in creative ways to make the game do things in a different way. (ATG Rockets in RHS come to mind)

One warning though: after you spend hours of your precious leisure time creating your mod, be prepared to have it ripped to pieces by the uninformed masses. (Oh, BTW, the pay for modding the game is the same as the Matrix team gets)

I am not just saying this to be combative. I truly mean it: if the game is not working properly then it will be the end users who will most likely be the ones to fix it. You got a problem? Man up and lets get it fixed.

(edit to clean up typos and add comment about manipulating the DB)

[Bullwinkle58]

Well said.

[John 3rd]

Vettim--Nicely done Sir.

Back in WitP we had some discussions regarding the Japanese Circle Four program. We bantered back-and-forth some and a lot of notes were taken. AE came upon the scene and those notes were put to use. Several "minions/friends" jumped in and the scenario became Reluctant Admiral. The 'finished' product was far better then what was envisioned! The process only took a couple of years or so from ideas to completion.

I am computer illiterate and came to understand and love the Scenario Editor. IT is made for idiots (that's me) and anyone can put something together. Word of Warning (like Vettim) it is frustrating and time-consuming...

Good Luck!

[Nomad]

So Tim, have you started rounding up your team yet? [:)]

[nashvillen]

+1

[vettim89]

ORIGINAL: Nomad

So Tim, have you started rounding up your team yet? [:)]

Not really but I have finally started testing. I believe in data making the argument not impressions. That means very boring and repititive testing. I ran 90 turns today doing some preliminary ASW research. Problem is I think 90 turns may not be enough. I may need even more iterations. Anybody know a good statistician?

[PaxMondo]

I'm not bad.  Issue is that with so many variables it isn't easy to give concrete answers.  Or better to say, I can give a concrete answer using the number of variables involved and conservative assumptions as to their respective independency.  The number of runs to get even mediocre variances runs into the thousand's quickly due to Gary's fundamental belief in random factors and the use of so many variables.  Even if you had the formula's in front of you to exactly assess the relative independence of each variable, there are still so many variables that it would make statistically significant run numbers problematic.  =>  Think hundreds of runs on each variable to get only 1 or 2 sigma type results.  Worse, we know that some of the formulas have boundary condition issues: they are not normally distributed (they are single sided distributions, or worse truncated).  This greatly complicates the equations.  For people who think this doesn't matter, I remind you of Challenger.  NASA spent billions on non-normal statistics after Challenger.  Money, I might add, that was well spent as it is used in almost every industry today from electronics to pharmacueticals.
 
So, bottom line: I might run 10 - 20 tests on maybe 3-4 values for each variable.  Statistically significant?  Nope.  Useful info?  Yep.  Sometimes great compromise is the only way to get a project done.  You are still talking hundreds, if not thousands, of runs ... bless you for even thinking of doing it ... [;)]
 

[vettim89]

ORIGINAL: PaxMondo

I'm not bad.  Issue is that with so many variables it isn't easy to give concrete answers.  Or better to say, I can give a concrete answer using the number of variables involved and conservative assumptions as to their respective independency.  The number of runs to get even mediocre variances runs into the thousand's quickly due to Gary's fundamental belief in random factors and the use of so many variables.  Even if you had the formula's in front of you to exactly assess the relative independence of each variable, there are still so many variables that it would make statistically significant run numbers problematic.  =>  Think hundreds of runs on each variable to get only 1 or 2 sigma type results.  Worse, we know that some of the formulas have boundary condition issues: they are not normally distributed (they are single sided distributions, or worse truncated).  This greatly complicates the equations.  For people who think this doesn't matter, I remind you of Challenger.  NASA spent billions on non-normal statistics after Challenger.  Money, I might add, that was well spent as it is used in almost every industry today from electronics to pharmacueticals.

So, bottom line: I might run 10 - 20 tests on maybe 3-4 values for each variable.  Statistically significant?  Nope.  Useful info?  Yep.  Sometimes great compromise is the only way to get a project done.  You are still talking hundreds, if not thousands, of runs ... bless you for even thinking of doing it ... [;)]

Well that's the rub now isn't it? There are a considerable number of Forum posts regarding one facet of the game or another being broke. By that I think people feel the results they are seeing are either ahistorical or unrealistic or highly improbable. The problems is that many times one player may see something while others don't. Why is that? Well I think sometimes players place orders in a particular way that introduces one of the non-linear formulas into the calculations. In other words they are inadvertantly creating the odd results

My small example was trying to look at ASW. I was testing the effect of both number of ships in an ASW TF and the actual ASW values. For example: is there a significant difference between two ships with ASW value of four each and four ships with ASW value of two each. I was also trying to get an idea of what detection values for both the subs and the ASW TF would be. Well, I ran 30 turns of three different iterations. Problem I had is one ASW TF ALWAYS had a detection level of 4 while the other one had a detection level of 0 90% of the time. Well I obvious found a variable somewhere to see such a out of whack result. That is unless you believe I can flip a coin 90 times and get head 88 of them. Yes, possible but still unlikely. Ergo, back to the drawing board. The only thing I can think of coming into play that I did not make the same is the leadership values for the commanders of the USN DD's and the IJN SUbs

[Mac Linehan]

ORIGINAL: vettim89

ORIGINAL: PaxMondo

I'm not bad.  Issue is that with so many variables it isn't easy to give concrete answers.  Or better to say, I can give a concrete answer using the number of variables involved and conservative assumptions as to their respective independency.  The number of runs to get even mediocre variances runs into the thousand's quickly due to Gary's fundamental belief in random factors and the use of so many variables.  Even if you had the formula's in front of you to exactly assess the relative independence of each variable, there are still so many variables that it would make statistically significant run numbers problematic.  =>  Think hundreds of runs on each variable to get only 1 or 2 sigma type results.  Worse, we know that some of the formulas have boundary condition issues: they are not normally distributed (they are single sided distributions, or worse truncated).  This greatly complicates the equations.  For people who think this doesn't matter, I remind you of Challenger.  NASA spent billions on non-normal statistics after Challenger.  Money, I might add, that was well spent as it is used in almost every industry today from electronics to pharmacueticals.

So, bottom line: I might run 10 - 20 tests on maybe 3-4 values for each variable.  Statistically significant?  Nope.  Useful info?  Yep.  Sometimes great compromise is the only way to get a project done.  You are still talking hundreds, if not thousands, of runs ... bless you for even thinking of doing it ... [;)]

Well that's the rub now isn't it? There are a considerable number of Forum posts regarding one facet of the game or another being broke. By that I think people feel the results they are seeing are either ahistorical or unrealistic or highly improbable. The problems is that many times one player may see something while others don't. Why is that? Well I think sometimes players place orders in a particular way that introduces one of the non-linear formulas into the calculations. In other words they are inadvertantly creating the odd results
My small example was trying to look at ASW. I was testing the effect of both number of ships in an ASW TF and the actual ASW values. For example: is there a significant difference between two ships with ASW value of four each and four ships with ASW value of two each. I was also trying to get an idea of what detection values for both the subs and the ASW TF would be. Well, I ran 30 turns of three different iterations. Problem I had is one ASW TF ALWAYS had a detection level of 4 while the other one had a detection level of 0 90% of the time. Well I obvious found a variable somewhere to see such a out of whack result. That is unless you believe I can flip a coin 90 times and get head 88 of them. Yes, possible but still unlikely. Ergo, back to the drawing board. The only thing I can think of coming into play that I did not make the same is the leadership values for the commanders of the USN DD's and the IJN SUbs

vettim89 - Gents,

Am following this discussion with interest. Your above points are food for thought. I do not yet have the sheer playing / observation experience that you do, so will try to learn as much as possible from the remarks in this thread.

What an awesome game!

Mac

[PaxMondo]

ORIGINAL: vettim89


Well that's the rub now isn't it? There are a considerable number of Forum posts regarding one facet of the game or another being broke. By that I think people feel the results they are seeing are either ahistorical or unrealistic or highly improbable. The problems is that many times one player may see something while others don't. Why is that? Well I think sometimes players place orders in a particular way that introduces one of the non-linear formulas into the calculations. In other words they are inadvertantly creating the odd results

I agree with you 100%.

I read it routinely. Someone has an outcome that they feel is impossible, so the engine is broke. Wrong. Mao did escape in reality against impossible odds. In combat, the impossible happens fairly routinely. However, on average, we know what an outcome should be if we know enough about the forces involved.


Simple example: fight between 2 men, one with ak-47, the other with a knife at 10 feet distance. Odds are with the ak-47, Right? But to say it is impossible for the man with the knife to win isn't accurate either ... This, I beleive, is Gary's fundamental premise in all of his games ... chance is a big factor. And it can come from many areas. Using my example, some obvious areas for luck would be: ak-47 jams, good throw, ak-47 hesitates, ... What are the individual odds of each of these? Low. The cumulative odds? Still low, but no longer undetectable. Fact is, we know that the guy with the knife does have a chance ... but we'd all stll prefer to have the ak-47 though.

Sounds a lot like WitP, right? IJ is doomed from the start ... we all know this. BUT, look at C&G's or Pzb's AAR's. Little bit of luck at the right time combined with good tactics and strategy and you can pull off BIG upsets. Have they won? No, but they are doing far better than historically. Flip the coin to Cuttlefish's AAR ... little bit less luck and your CV's react into disaster.

Have I seen some weird things happen in WitP? Oh yeah. Like when my 6 DD's got ambushed by 4 PT's and got clobbered (I lost 2 DD's and didn't even scratch a PT's paint). Does that mean the engine is broke? Hardly. I've been places I really, really shouldn't have been before and I'm still here and I know the odds of that outcome were really low.

PS: Remember if you run a test, you have to exit WitP each time completely to run the next one. The game's random number generator counters won't reset unless you do. Adds considerable time to each test run.

[dwg]

on average, we know what an outcome should be if we know enough about the forces involved. Simple example: fight between 2 men, one with ak-47, the other with a knife at 10 feet distance. Odds are with the ak-47, Right? But to say it is impossible for the man with the knife to win isn't accurate either ...

Which is an interesting example, but for slightly different reasons than intended. It's perceived wisdom that gun has an advantage over knife, but do some googling and there turns out to be a fairly significant number of people, largely amongst those with some actual exposure to knives, who argue that the knife has the advantage inside a certain distance, 21ft being often quoted. So many people will look at a sim of knife vs gun that has the knife win and say the algorithm is wrong, a few will look at it and say that statistics must let the knife win occasionally, and a very few will wonder which model the algorithm is using, and whether it's a statistical fluke, or a model different to the one people expect, but based on valid data.

Most people playing WitP aren't weapon historians, they may have some knowledge, in some cases very detailed knowledge, of the Pacific war, but you really need compiled results from combat performance to understand how a weapon is actually performing. Hedgehog is a good example of this literally extending onto the battlefield, captains initially didn't like Hedgehog, it didn't have the visual effects of depth charges, and they'd often hold off on Hedgehog until they had exhausted their other ASW armament. But then Hedgehog started sinking submarines, lots of submarines, and they realised their perception that Hedgehog wasn't very good was actually completely wrong. (History paraphrased from Brown's 'Atlantic Escorts'). If the people using them could get the wrong impression of their effectiveness it's hardly surprising the same may be true of non-specialists 65 years later. As another example, 3" rockets had a considerable reputation hung under Typhoons, but don't seem to have been the tank-killers of repute, lacking the accuracy to be anything but an area weapon -- the OR review of Falaise Gap found a grand total of 17 vehicles had actually had a direct hit, vs the hundreds claimed. (Which isn't to say the 3" RP wasn't a devastating weapon both at land and at sea, it just wasn't an individual tank killer).

What this means for WITP AE is that many enthusiasts are going to be making assessments based on the public face of weapons, while the designers will have gone a level deeper into the history (within the limits of the original algorithms), and the real historical performance may have been different again. These differences in perception inevitably lead to some friction, the difficulty comes in knowing whether the person commenting has the background to make his comments something worth listening to. So if you think something is wrong, bring data and a sound argument, not just an opinion!

[witpqs]

A very astute point, dwg.

[SuluSea]

ORIGINAL: dwg

Which is an interesting example, but for slightly different reasons than intended. It's perceived wisdom that gun has an advantage over knife, but do some googling and there turns out to be a fairly significant number of people, largely amongst those with some actual exposure to knives, who argue that the knife has the advantage inside a certain distance, 21ft being often quoted. So many people will look at a sim of knife vs gun that has the knife win and say the algorithm is wrong, a few will look at it and say that statistics must let the knife win occasionally, and a very few will wonder which model the algorithm is using, and whether it's a statistical fluke, or a model different to the one people expect, but based on valid data.

Most people playing WitP aren't weapon historians, they may have some knowledge, in some cases very detailed knowledge, of the Pacific war, but you really need compiled results from combat performance to understand how a weapon is actually performing. Hedgehog is a good example of this literally extending onto the battlefield, captains initially didn't like Hedgehog, it didn't have the visual effects of depth charges, and they'd often hold off on Hedgehog until they had exhausted their other ASW armament. But then Hedgehog started sinking submarines, lots of submarines, and they realised their perception that Hedgehog wasn't very good was actually completely wrong. (History paraphrased from Brown's 'Atlantic Escorts'). If the people using them could get the wrong impression of their effectiveness it's hardly surprising the same may be true of non-specialists 65 years later. As another example, 3" rockets had a considerable reputation hung under Typhoons, but don't seem to have been the tank-killers of repute, lacking the accuracy to be anything but an area weapon -- the OR review of Falaise Gap found a grand total of 17 vehicles had actually had a direct hit, vs the hundreds claimed. (Which isn't to say the 3" RP wasn't a devastating weapon both at land and at sea, it just wasn't an individual tank killer).

What this means for WITP AE is that many enthusiasts are going to be making assessments based on the public face of weapons, while the designers will have gone a level deeper into the history (within the limits of the original algorithms), and the real historical performance may have been different again. These differences in perception inevitably lead to some friction, the difficulty comes in knowing whether the person commenting has the background to make his comments something worth listening to. So if you think something is wrong, bring data and a sound argument, not just an opinion!

Simply outstanding! [&o][:)]

[Rainer]

I sincerely wish and hope a lot of the "game's broken/borked" people have a chance to read this, including the fraction of people who, after reading one or two books about the events, believe they are experts or even "historians".

After playing GG's games including PacWar, WitP and Witp/AE since the early 90s I am convinced that he and his team brilliantly suceeded in producing war games which give players a chance to experience the thrill, excitement and horror of those days.

Those who insist on re-creating history should watch movies or read books (and play Witp/AE as a welcomed diversion [;)])

[JWE]

Oh, yeah! This game sucks so bad, in every dimension, one wonders why several thousands play it and have a ball doing so, while 5 or 10 "experts" are vomiting "I didn't have anything to do with it, so I hate it!" bile.

Reminds me of a thing I just got in a fortune cookie; "If you must question it, you are likely wrong." Yep, shit happens and, Zen notwithstanding, it happens on your windshield, your opponents windshield and, frankly, on your mother-in-law's windshield while she's watching Wheel of Fortune. Doo-doo spreads, and it spreads judiciously; and it spreads in accord with an alghorithm that only Gary Grigsby truly understands ( no matter how much the rest of us really want to play).

All I can say is (humbly, gently, and oh, so graciously) bite my Willie the Weenie, and I hope to God if it ever gets down to cases, that I face people like that. They will very likely allow me some sleep.

[Rainer]

I'm with you.

My concerns are:
- Members of the Voluntary Dev Team give up in frustration (that already happened as we all know)
- other members of the Dev Team believe the few nitpickers have to be pleased and try to improve the product in the wrong direction (happened as well)

I wished the Silent Majority sometimes would not be so silent.
Oh well ...

[oldman45]

I know that the hardcire players will cringe at this, but I want the game to "feel" right. I know its not perfect and things will happen from time to time that make an eyebrow go up but if at the end of the day the rounds I played "felt" right then I am a happy camper. When I see many of the examples that people post, I have to say, I have never seen them in my games. Of course I only play the AI because its human nature to want "big" attacks and those seem to screw up the engine faster than anything else. After the changes to AAA and subs torps that the babes team did, the only thing that still causes that "doesn't feel" right moment is ship combat. Something about it just doesn't seem right in many of the capital ship actions I have. I am one happy camper with this product and the support it receives. One day I might even figure out how to get my logistics "tail" right [8D]

[stuman]

ORIGINAL: Rainer

I sincerely wish and hope a lot of the "game's broken/borked" people have a chance to read this, including the fraction of people who, after reading one or two books about the events, believe they are experts or even "historians".

After playing GG's games including PacWar, WitP and Witp/AE since the early 90s I am convinced that he and his team brilliantly suceeded in producing war games which give players a chance to experience the thrill, excitement and horror of those days.

Those who insist on re-creating history should watch movies or read books (and play Witp/AE as a welcomed diversion [;)])

But what if I am a whiney, self absorbed, immature, smug, know-it-all touble maker who never got picked to play on sports teams. Don't I have rights ! Don't I get to f**k up the game for the rest of you ?

I think I need a hug.