Using the battlecalculator when attacking (moscow) or any other large battle

[oysteilo]

If moderator feels this belongs elsewhere please move

In a recent game of global (dice) Germany decided to attack Moscow with a bunch of units (about140-150 units). From the battlecalculator Germany is a 96% favorite and on average should have 33 units remaing. The actual battle goes poorly for Germany leaving them 8  bombers and an armour. The German player feels this is really poor dice. But is it really? The battlecalculator also says it is a 4% chance for Russia to win and they should on avarage have 10 or 11 units left. The question is what does these % mean? The Russian % is not that interesting, but what does it mean for Germany? Can you assume the outcome is a normal distribution centered around 33 units remaing? In that case it is equally likely to have 23 or 43 units remaing? Does it also mean it is a 96% chance for Germany to have from 1 – 66 units remaing (assuming 67+ units remaing is insignificant)? If you can look at it this way it also means it is a 48% chance for Germany to have 33 or more units remaing and it is a 52% chance to have 32 or less units remaing. Maybe the actual outcome of 9 units remaing is not that unlikely? Is there a good way to estimate this? I am estimating about 15% chance for the actual outcome, but I have no scientific data to back this. Another way to look at this is if Germany throws a slighly poor dice in round one and Russia throws a slightly good dice in round 1 it is easily a change of 10-15 units. If everything goes average after this the outcome is by far no longer 33 units. Comments anyone?

[WILD BILL]

I’m not a big numbers guy, but I will run a battle calculator as many do to get odds on battles. I don’t really count heavily on it because as you pointed out there is a pretty big grey area when it comes to dice. If I have 90% odds, then I would expect to win the battle, but I wouldn’t count on having the units shown when its over (or units equivalent to the IPC value given). I think you hit the nail on the head with the opening round setting the tone. I would even point out that the AA round could change the battle quite significantly too. Say Russia can get off 12 AA shots (4 AA guns) and kills 3-4 air units instead of the expected 2 planes. That alone could move the scale 5-10% which could end up costing Germany 10+ units over a 5-6 round battle.

What might give you more insight is to play out the first round, then run the battle calculator again with the units still in play (no AA guns to contend with). If as you say Germany under performed (by say 5-7 hits), but Russia throws average, slightly above average or hits a couple extra AA shots it could move the scale from 96% to 75%. Germany would still most likely come out on top, but instead of expected 33 units surviving, you’re probably looking closer to 20 units give or take.

[JDOW]

Can you assume the outcome is a normal distribution centered around 33 units remaing? In that case it is equally likely to have 23 or 43 units remaing?

This is a good question. I’d say no, it is not a normal distribution. It depends on the units composition, so I’d say that distribution depends on each battle.
In any case it IS a distribution arount the average result, however I guess it is not a bell curve.

Does it also mean it is a 96% chance for Germany to have from 1 – 66 units remaing (assuming 67+ units remaing is insignificant)

Yes! And this is the most common mistake many A&A players do when calculating battles. They only see a percentage of winning. However that also includes pyhrrus victories which unacceptable losses.

Maybe the actual outcome of 9 units remaing is not that unlikely? Is there a good way to estimate this? I am estimating about 15% chance for the actual outcome, but I have no scientific data to back this

Good point. I think Bottom 15% is a good estimate. I think gamerman has a method to approximate this more precisely.
The key is to use a different battle calculator: http://www.dskelly.com/misc/aa/aasim.html

I think this one allows better calculations than the tripleA calculator. but frankly, I am too lazy to made myself familiar with the other one and only use the tripleA calculator.

Cheers,
Tobias

[JDOW]

Try the calculator I linked in my previous post. I guess the pie, bar and loss chart could answer your questions.

[ItIsILeClerc]

http://www.dskelly.com/misc/aa/aasim.html

Great calculator!
Well, I guess it is more about the functions of it.

In TripleA games (which I do not play very often) I’d run it maybe multiple times for every single crucial battle, playing with some different “order of losses”. This can be very crucial. For example: the BC automatically assumes that STRat bombers in a defending army are taken as losses first (all of them). Changing the OOL into preserving the STR (more or less) will give very different results!
Little sidenote: I found out that the TripleA BC gives wrong results if simply changing the OOL (at least in this particular example). For a more correct ‘prediction’, I save a copy of the game and fight the battle through, taking all the losses myself from round to round to see the end results.

In FtF games around the table we never used a BC untill our playstyle changed. Instead of many smaller battles we now have fewer but much larger battles, where it is impossible to predict who will win. But we use a simple (excelsheet) BC based on pure LL only, just to have a rough clue about where the battle is going.
We found this necessary because in such large (crucial) battles, simply adding up attack factors versus defense factors isn’t accurate enough (for us). Also, comparing number of units (just a little more complex) is unsatisfactory because attacking with 100 vs 98 might give the attacker 90% chance of victory, but attacking with 1 less unit (99 vs 98) may drop the chance to about 50% and how can we predict where this treshold is? To make matters even more complex, the treshold shifts with unit composition too…
So in a FtF game, we usually use the (simple) BC for ‘Moscow’, ‘London’, ‘Midway’ or ‘Normandy’ battles. Sometimes the ‘Egypt’ battle as well, if it is applicable.
Winning or loosing these type of battles usually means winning or loosing the game, so why would we want to engage without a clue about who will win it  :wink:.

[ChocolatePancake]

I always battle calc any battle that isn’t immediately obvious, and I expect my opponents to do the same if they want.
In my experience it takes just as long to add up all the attack power vs defense power as it does to simply plug everything into the calculator, and the calculator is far more accurate.
I’ve never seen it take more than an extra, maybe, 5-10 minutes over the course of an all day long game, and I find it makes the game better overall as it helps people make less stupid mistakes.

[Zhukov44]

A google search shows the original AACalc (favored by a&a geeks in the Revised days) has now migrated to this site.

http://calc.axisandallies.org/

Unfortunately, both the above calc and the DsKelley calc require upgrades to be useful to global players.

[Witt]

I never use the calculator.

[gamerman01]

Axisandallies.org has a pretty good calculator - look on the ribbon of home/help/search/profile/my messages/members/AACALC/logout

I’ve heard some people don’t have this on their ribbon for some reason.
Oh, Zhukov gave you the direct link to it.

You do have to modify it a little for G40.  If there are 2 AA guns and >5 planes flying over, I just subtract a fighter and don’t check the AA gun box.  I enter AA guns as bombers (along with converting that number of infantry to “bombers”) to approximate.  Be sure to change the order of loss to put “bombers” first, and delete the other “bom” entry in the defensive OOL.

You’ll have to approximate for things like tactical bombers too, of course.  But the site’s calculator gives you a complete distribution of results as well as highlighting means, medians, and 1 or 2 standard deviations above or below.

I use another, older, calculator that is quirky but has some unique features and also gives a complete probability distribution.  I want to know what the chance is that I start losing aircraft to take a territory, for example.  Triple A calc is quite limited by not giving you distributions.  Don’t ask for it - even getting it set up and working is quirky.  Between these two calculators (the one on the site is sufficient, I think) I have a very clear idea of what I’m getting into and what the chances are.

[gamerman01]

If you want to fully understand the AACALC on this site (see link from Zhukov), you need to know the rules to the older game (Zhukov said it was Revised).

One big difference is the AA guns shot separately at fighters and bombers, so could directly hit bombers, whereas in G40 the attacker can choose his losses and shield his bombers.  Also, there was no limit to AA firing like there is in G40, so if there’s one AA gun and you’re flying 5 fighters over, the calc won’t do that correctly.  You’ll have to run it with no AA gun, with 5 fighters, then with 4 to get an idea (even though there’s a 1/216 chance that 3 fighters are downed).  Or run it with the AA gun and know that your actual chances are a bit better than it will show (because the AA cannot fire 5 times like it used to)

[oysteilo]

Thanks a lot gamerman01 (and all others) . I like the original calculator on the site because of the distributions but find it hard to use for global games. I will take your suggestions though and play around with a little bit. I think it is critical to understand how likely it is to win a battle but loose too much!

[gamerman01]

Another tip - I usually put in double the number of carriers since they are 2-hit in G40.
I understand that this overstates the attack power (defense power too) of the carriers, but it’s not a lot.  Maybe I’ll compensate by changing a fighter to a cruiser or something.  It still beats TripleA calc.

Often there are some attacking and defending carriers, so it balances out.  (Calc is using 1-hit carriers that attack on a 1 and defend on a 2)

I’m just happy if I can help enhance someone’s game playing experience, so thanks for the feedback oysteilo

[baron Münchhausen]

@Gamerman01:

Another tip - I usually put in double the number of carriers since they are 2-hit in G40.
I understand that this overstates the attack power (defense power too) of the carriers, but it’s not a lot.  Maybe I’ll compensate by changing a fighter to a cruiser or something.  It still beats TripleA calc.

Often there are some attacking and defending carriers, so it balances out.  (Calc is using 1-hit carriers that attack on a 1 and defend on a 2)

I’m just happy if I can help enhance someone’s game playing experience, so thanks for the feedback oysteilo

To accurately simulate the G40 2-hits Carrier A0 D2 C16: use 2 transports on offence, or 1 transport and 1 carrier on defence, and put their casualty order before any combat unit, according to your taste and battle scenario.
Each transport counts as 1 hit casualty with no attack or defence value.

[gamerman01]

That’s a good tip - my TUV calcs won’t be off as much (and the win/loss results will be more accurate) - I forgot transports used to take hits and could be taken first, since AA50 came out in the fall of 2008  :-)

[calvinhobbesliker]

Wait, doesn’t the dskelly calculator have a Global 1940 version? It’s A3 (Alpha 3?) instead of 2nd edition, but that just means the only issue is AAA guns, right?

[gamerman01]

@ShadowHAwk:

This will actualy affect the outcome of your battles a lot and your calculation will be off.

2 inf + 1 bomber have the same attack value as 3 art. But if putted against 3 inf the 2 inf + bombers comby has a slightly higher chance to win.

Of course, because you’re losing 2 attack power with the first two casualties instead of only 1, and the 4 remains until last instead of a 2.

First of all, my approximations with carriers or tacs are done normally in much larger battles than 3 on 3 and the error is in fact negligible.  Again, the point is to get a probability distribution which Triple A doesn’t provide.  None of the good calculators I’ve found have been updated to accommodate G40.

My approximations using bombers in place of infantry/AAA is pretty accurate, because in almost all cases those units will be eliminated in the first round of the battle.