Prerequisite for Dice Complainers

[mmd1]

i know the odds on that 6 inf attacking 2 inf, 3 ftrs utzchi[s - a 5% outcome is not what is remarkable. what is remarkable is doing an attack with a 5% chance of success in the first place. my record of inf missing is 23, but i have had plenty more around 17, 18 etc. that have also missed. these are not so remarkable but it reinforces the mentality that attacking inf are there to die instead of other units. when i have a lot of inf in an attacking stack i subtract 10-20% from the odds before deciding whether it is worth doing. ok, the payoff is potentially very big in the manchuria attack i describe. fair enough if that is your thing i guess.

i accept that small attacks involving between 10 and 20 units in total can go very badly wrong on occasions. with smaller attacks almost anything goes. but how does the result you describe with rem occur? i'd like some more details on that. surely, the larger the stack is the more likely you are to get something closer to average. if your big stack is 40+, how does a stack of mostly inf get an 85% hit rate. that does suggest something is pretty funky to me.

[UtzChips]

here's the only things that is bugging me about this exchange:

the chance of devastating defeat in SE in lukalion's example was 1/300, NOT 1/5000.

Not understanding probabilities (Utz here - not lukalion's point) is a good first step to totally over-stating how often your game will be "ruined" by the dice, or in picking a level of conservatism you prefer for your moves.

If you really want to ruin a game, have Germany sealion UK on G1. in 1/33 games, you'll take Berlin (and even probably win... . Of course, you have to ruin 32/33 of your own gaming experiences to enjoy this atrocity.

I don't understand how you get 1 in 300.

If a 2% chance of occurrence out of 100 is: 100/2 = 50 thus 1 in 50 or 49:1
Then a .02% [2/100s of 1%] chance of occurrence out of 100 is: 100/.02 = 5000 or 4999:1

So what am I doing wrong?

If you go to http://www.axisandallies.org/aacalc and put in the units, you will find that U.S.A. has 1 inf left either .01% or .02% of the time, depending on what it averages it out to. Sometimes you'll get .01%, other times .02%

That is 2/100s of 1%.

mmd1,

You will have to talk to mmd1 about that battle in Berlin that we had. It was a pbe game & I don't remember exactly how many I had.

Here is what I do know:

*Russia can hit WR with 9 inf 1 art & 2 tanks [punch of 18] and roll all white the 1st roll & that is expected to occur [as far as I am concerned] because there is a 2.76% chance of occurrence.

*Conservative/optimal play will win you a higher percentage of games over the long run.

*I am of the understanding that Aeroforce didn't like to get involved in even semi-major battles without a more than 90% chance of success and he had a win rate of 90%+.

What I did not know:

Is that one needs to have a "Plan B" for when an 1/200s of 1% occurrence happens.

[bmaster]

That is 2/100s of 1%.

No. that is 2/100 of 100%. Or 2/100. However, having 1 infantry left on defense was irrelevant. what matters is the odds the attacker fails to take SE, as then usa can build there. That's the 1/300 shot for the strategically bad outcome.

... per the 6 inf attacking scenario, yeah the attacker was probably deluded about his odds. If he was badly losing the game, there could be some logic to it, but basically it's a bad bet.

[UtzChips]

bmaster,

We are, for some reason, miscommunicating. Maybe you think I took 2%, converted it to a decimal [.02] and left the percentage sign there? I did not.

2% in decimal form is .02

I am not talking about 2%. I am talking about .02%, which in decimal form is .0002

2/100 is 2%. See:

http://en.wikipedia.org/wiki/Percentage

If you use the battle calculator at http://www.axisandallies.org/aacalc and have 3 inf 1 art & 2 arm attack 2 inf, it shows that the attacker will be left with 1art & 2 tanks 2.09% of the time, 2 tanks 0.23% of the time & 1 tank .05% of the time.

EDIT: In Lukalion's example, the odds of U.S.A. having 1 inf left is immaterial. Germany needs to have 1 tank left to be successful, and that occurs at least 99.98% of the time, since the battle calculator shows Germany being left with AT LEAST a tank 99.98% of the time.

I get that 99.98% by running the calc using the "10,000x times" choice. If you use it, and run it several times, you should get an result where it shows you have at least one tank left .04% of the time. You may have to run it at 1000x then 5000x & then 10k because the margin of error is so close.

Then I add up the percentages for all the different possible outcomes for Germany and it totals 99.98%.

[mmd1]

am hitting personal bests at an astonishing rate - just lost a 99.8% battle. attacked 6 sodding inf again with 9 inf, 1 tnk, 2trs!!! and one defending inf survived 0.1% against 1 russian ftr. inf hit 4/6, 2/6 (yes that's right i missed everything, but that happens so often it barely rates a mention) and then 4/4. viola.

oh and another 4.2% battle against japan in manchuria came up roses against me. it's all the rage. and why not when it works so well at gto.

next time the server is down for maintenance try smashing it a few times with a sledgehammer please.

[bmaster]

Oh hey utz sorry I think I misread lukalions post. I thought it was three inf two tanks attacking two inf. Now I notice he mentioned ",art" as well, so I assume that's 1-2 art as well which likely does make the odds much better as u describe (not at my pc so can't check)

[UtzChips]

Oh hey utz sorry I think I misread lukalions post. I thought it was three inf two tanks attacking two inf. Now I notice he mentioned ",art" as well, so I assume that's 1-2 art as well which likely does make the odds much better as u describe (not at my pc so can't check)

It was 3 inf 1 art 2 tanks v. 2 inf. So, without the art, my calc shows that Germany having at least a tank left approx. 99.68% of the time.
100/.32 = 312.5 or 311.5:1 & I have close to the 300:1 probability you quoted!
Problem solved!
So, I thought to myself: "Self, if that's true, surely adding 1 art isn't going to raise the odds all the way to 4999:1!"

Then I realized, USA doesn't need an infantry left, it just needs to kill all German units.
So I ran the calc again & Germany has at least a tank left approx. 99.96%, so approx .04% of the time they lose it all.
100/.04 = 2500, or 2499:1 success rate for Germany with 3inf 1art 2tanks.

[bmaster]

makes sense. nice to meet somewhere in the middle

[rockismyreligion]

makes sense. nice to meet somewhere in the middle

3s and 4s reflect the middle too.

[UtzChips]

makes sense. nice to meet somewhere in the middle

3s and 4s reflect the middle too.

More junk mail from U know who!

If you're gonna' try to be humorous, at least put a lil' effort into it! Please!

I mean, come on man! How many people do think appreciated opening up their e-mail to find your above response?

Was it really that funny?!

[rockismyreligion]

makes sense. nice to meet somewhere in the middle

3s and 4s reflect the middle too.

More junk mail from U know who!

If you're gonna' try to be humorous, at least put a lil' effort into it! Please!

I mean, come on man! How many people do think appreciated opening up their e-mail to find your above response?

Was it really that funny?!

He who slings mud generally loses ground.

[BKBrawler]

Utz and rock, you guys are something else, really.

On another note: had a great game going a little while ago with a very good opponent. Played as Axis, opponent was Allies, and I was stacking egypt hard. Definitely at least 7 or 8 rounds in and he still couldn't have hit it with better than 30% odds or so. The board was getting stagnant for the Allies, Africa was blue, Japan had 3 ICs and 9-10 transports and was trading in Kaz and Novo. Germany was well stacked too with all its tanks and planes, with stacks in Berlin and EE, trading WE with a couple of UK inf a round, threatening a Cau stack the minute the Allies peeled off to bolster Russia or try and push back into Asia.

So my opponent decided to mix it up to try and break the deadlock (which was slowly tilting Axis favor, I think its fair to say), so he takes SE with US- 10 units, loses 2, 8 left. His fleet off SE consisted of 1 carrier, 1 battleship, 2 fighters, 2 destroyers, a sub, and 5 transports. He has 4 transports also in 12. No other boats. He has a couple of fighters and a bomber in Cau, and another 2 fighters and 2 bombers in Cau for UK (I think).

On Germany's turn I retake SE.

When it comes around to Japan's turn, I hit it with everything that can reach, thinking to land a decisive blow by wiping out the US fleet, and hoping to have enough units left to seriously threaten Atlantic. I think conservatively he would have been forced to stay off Africa and Europe for at least 2-3 rounds to rebuild a viable escort fleet and 2nd half of transport chain. That felt like it would be a pretty bad setback for the Allies. So in I go with 2 battleships, 1 destroyer, 2 carriers, 3 transports, 6 fighters, 1 bomber. I ran it on the calculator first (as I usually do for big battles) and came up with these results:

99.2% chance to win, average IPC loss for me is 53.735.

So on average, I'd wipe the fleet and lose the transports, destroyer, and a couple of fighters. Wow, that's a nice trade. But even if it goes poorly, it seems I can reasonably be expected to survive at least the battleships, a carrier, and a couple of fighters. That's just fine with me. Game on.

First round of combat his sub hits. Then I roll...

1 hit.

Umm, yikes!

He defends 6.

He didn't lose a single unit and I lose half the fleet and a fighter. Just like that. Game over even with a withdrawal as he has enough units to take out the fleet with what he has. Already knowing I'm done I play it out and the two sides completely annihilate each other.

It was a very fun game, and this battle doesn't change that, but I wonder why the game should have been ended like that. I don't think its fair to say I would have won without 14 going poorly, but I think its fair to say I liked my chances without the dicing. Seems a Middle Luck option would be nice when you get results like that, but dice are dice, and no big deal. I'm sure they'll come around for me in a battle one day. Or will they?

One thing I realized is that I rarely "give" dicings. Is this because I get unfair dice more often? Nope! Actually there's a much more rational reason. I play a conservative style for the most part. There are certainly calculated risks I'm willing to take throughout a game, usually early on in small battles when there is more variation yet they can mean a lot, but as the game progresses I definitely get much more conservative, playing the IPC and position game to try and keep squeezing an extra unit out of my builds and taking from the other guy, to build up to the point the percentages slowly work up for me, and once that starts happening I won't attack until I start seeing odds rising to 85%+. But before that happens, I simply wouldn't risk large numbers of units or expensive ones by placing them in positions to get hit by overwhelming force or have low chance to win. So the opportunity to "give" dicings almost never happens for me as a result of my play style.

So for anyone else who feels they get diced more often then give it, that's probably due to your play style, which I think is a positive. You can't give a dicing unless you put big units in harms way for a bad battle.

That's my shpeel for the day.

He's a talker that BK!

[bmaster]

fun story, bk. i think that scenario's happened to all of us. It's hard to resist that killing blow opportunity with fleets.

That said, if you were gonna get a 1% battle outcome somewhere else, i bet your opponent could've just taken berlin with a "suicidal" uk drop.

A fun feature one might consider for non gto games:

1) give each player some number of "special" roll chits (say, 3 per game)
2) using a chit allows you and your opponent to get exactly average dice rolls (or LL results) in any one round of any one battle of the game, so long as you announce it before you roll.

I kind of like this idea because under no circumstance does it allow you to gain a huge advantage, and in fact it precludes you from the chance to get lucky as well as unlucky. But it does require some skill to pick those three 3 rounds of battle in the game that are so central to your strategy that you want at least average results.

On the whole, I'd still want to play with few "chits" per game, because too many and suddenly high risk high reward strategies are no longer an option for a player (like your opponent in that game) who is needing to make a risky move to survive. But having a few is a nice response to those concerned that games are only decided by luck.

So here's a question, if you had 3 chits, would you use one on ukraine R1 or Egypt G1?

[UtzChips]

BK!! OUCH !!

However, I get different numbers than you. If both Japan & U.S.A. take losses based upon the punch of the units [Japan taking losses on carriers after trannys & U.S.A. taking losses on carrier b4 fighters] I have you ending up with no units 2.2% of the time. That makes you a 44.45:1 favorite........nice.......but no guarantee with dice

Your units: 2 BBs 2 CVs 1 DD 6 ftrs 1 bmbr
USA: 1 BB 1 CV 2 DDs 1 sub 2 fighters

Since U.S.A. has Libya to land his fighters & is a heavy underdog, there would be no point in him trying to save his carrier & should have taken the loss on the carrier first.

[BKBrawler]

A fun feature one might consider for non gto games:

1) give each player some number of "special" roll chits (say, 3 per game)
2) using a chit allows you and your opponent to get exactly average dice rolls (or LL results) in any one round of any one battle of the game, so long as you announce it before you roll.

So here's a question, if you had 3 chits, would you use one on ukraine R1 or Egypt G1?

That's a pretty cool idea. I'd almost definitely use it R1, but in WR instead of Ukraine. I generally go heavy Ukraine, so the chances of failing are pretty low there, but WR is a bit more important in the sense that it has to win (and almost always does) but win it strongly, ie 6 or so inf surviving. If that were to go poorly, that's a tough hole to start in.

I don't think I would use one on G1. Even if you use it in Egypt, you have a good chance of only having 4 units survive, and possibly even 3, both of which are counterable by the UK, especially if it uses its own chit.

The chits are probably most useful for Japan to start. If its sub doesn't get sunk by the UK, it can use chits to hit Pearl, China, and Bury. Guaranteed success at these three juicy objectives for Japan would be pretty big.