13L Playoff Round2 hobo vs AllWeNeedIsLove

[hobo]

Average is much higher than 58.  You are not accounting for the transports lost.  Since the ool is slightly different, I calculated the losses per round and came out ahead in the battle by 2-3 hits so I should have lost the other tipped carrier and subs.  It’s not much different from the raw average survivors assuming carriers go first of +107 remaining or tipped battleships, 4tac, cruiser, and a fighter or 8 hits remaining.  I came out with 11 hits remaining.

[boldfresh]

So battle went to odds right hobo?

[JDOW]

@hobo:

Average is much higher than 58.  You are not accounting for the transports lost.  Since the ool is slightly different, I calculated the losses per round and came out ahead in the battle by 2-3 hits so I should have lost the other tipped carrier and subs.  It’s not much different from the raw average survivors assuming carriers go first of +107 remaining or tipped battleships, 4tac, cruiser, and a fighter or 8 hits remaining.  I came out with 11 hits remaining.

I am afraid I cannot follow. The +58 already took 3 TTs into account. I just put everything that was in that sz in the calculator and pressed the button. Why should the expected outcome be better than the Calculator calculation?
Sorry if I missed something btw :-)

[hobo]

maybe we are just looking at different things.  what I look at is the total def ipc value, or 211.  the average attacker loss is about 114.  subtract the two and that’s the average battle score assuming the attacker wins.  it’s close to 100.  the way I usually do things is to break the combat into rounds and calculate the hits.  from that you can get a close approximation of the surviving units.  in the case for this battle, I had 125 ipc worth left which is a little skewed due to the other tipped carrier I didn’t expect to have left.  in other words I loss 95 ipc worth, slightly less than the average.  add the carrier and a sub then it’s about average so again, I came out about 2-3 hits above average depending on how you want to look at things.  out of close to 70 dice rolled in the combat it’s not statistically far off the mean.  the results of the battle was about a 75-80% occurrence.

[JDOW]

@hobo:

maybe we are just looking at different things.  what I look at is the total def ipc value, or 211.  the average attacker loss is about 114.  subtract the two and that’s the average battle score assuming the attacker wins.  it’s close to 100.  the way I usually do things is to break the combat into rounds and calculate the hits.  from that you can get a close approximation of the surviving units.

Hey hobo, thanks for your explanation, I appreciate that.

Trust me, I do not want to annoy you or anything or I am just a fan of this game and I would to understand the calculation of battle outcomes in case I make a substantial mistake in so doing. Because I still do not understand.  :-(

Just a guess I have: If the calculator says the average outcome is +58 I am aware that this does not have to mean that +58 is the most likely outcome. Extreme example:A coinflip battle runs with 60% at +100 and 40% at -200 (nothing inbetween): Now the average is -20 although +100 is the most likely outcome (60%), so +100 is NOT a super big lucky swing but just happens 60% of the time. Is it this way you argue?

However I might still need to ask why you get a different average result than the calculator does. What does the calculator do wrong? Why is your average so much better even if you cannot just tip all carriers round 1?

Sorry again, I really don’t want to bother anybody or talk anybody into “Axis were so lucky”. I am just afraid I might have wrong thought processes or that I am missing something.

Cheers,
Tobias

[hobo]

not sure how I could explain it the way I just did.  all I can say is you are probably not looking at it the right way.  if you are just using the battle calc then just do this: 1) take the average attacker ipcs left, 2) subtract it from the starting attacker total(aka how much the attacker lost), then 3) subtract that from the starting total for the defender(since when the defender loses, it loses everything).  as I noted, that gets you close to 100 battle score.  I have NO idea where you get 58 or whatever.

the other way is to break the combat into rounds as I noted and you come out with similar results.  the reason is the average defender hits in the first round have NO effect on the total attacker combat power in the second round since it’s just tipping ships and losing 0 combat attack value carriers.  since the average combat length is 3 rounds, what you choose as losses, ie keeping 1 carrier vs doing default losses, really has little effect on the total outcome.

what happened in this battle, ie I came out ahead by 2-3 hits above average, was like a 70% occurrence if we just simplify each hit as over/under as about 5%, calculated compared to the total average dice rolled needed to conclude the combat(I went into this above).  that’s why the defender only has a 15% chance to win because they not only need a combination of attacker hitting below normal, but also the defender hitting above normal to a tally of 12 or so.

if you want to call this lucky go ahead as some of the people have commented when they compared a 33% occurrence for me with my italian bomber battle vs the 2% occurrence against me when his single ships shot down 3 planes and tipped my battleship.  interesting comparison.  to top that off, I nearly lost a 90% battle vs the USA fleet at gib the following turn when I should have had 3 surviving hits remaining on “average.”  combine the two, germany should have 6 extra air units.  just fyi, I will accept the this trade and lose the tipped japanese carrier and couple subs any day.  then again, apparently my success with games played here have been purely luck in the eyes of some.

[boldfresh]

Not much you say makes much sense but you are on a whole other plane hobo  :|

Still waiting for my revenge whenever you are ready.  :-)

[JDOW]

@hobo:

I have NO idea where you get 58 or whatever.

Just to clarify: +58 (actually 59 after 10k simulations) is the average TUV-Swing the battle calculator spits out. I attached a screen.

the other way is to break the combat into rounds as I noted and you come out with similar results.  the reason is the average defender hits in the first round have NO effect on the total attacker combat power in the second round since it’s just tipping ships and losing 0 combat attack value carriers.  since the average combat length is 3 rounds, what you choose as losses, ie keeping 1 carrier vs doing default losses, really has little effect on the total outcome.

what happened in this battle, ie I came out ahead by 2-3 hits above average, was like a 70% occurrence if we just simplify each hit as over/under as about 5%, calculated compared to the total average dice rolled needed to conclude the combat(I went into this above).  that’s why the defender only has a 15% chance to win because they not only need a combination of attacker hitting below normal, but also the defender hitting above normal to a tally of 12 or so.

the reason is the average defender hits in the first round have NO effect on the total attacker combat power in the second round since it’s just tipping ships and losing 0 combat attack value carriers.  since the average combat length is 3 rounds, what you choose as losses, ie keeping 1 carrier vs doing default losses, really has little effect on the total outcome.

Got you.

if you want to call this lucky go ahead

As I said, I do not call anything lucky. So far I just lived so far under the probably false assumption that an actual TUV score which is above the average TUV score by the calculator is a good result for the attacker. However I meanwhile have to acknowledge that in many battles an above average score is the most likely result.
Maybe also in this battle: You have 84% chance to win, if you win you have 7.14 units left which would equal e.g. 2 BB, 2 AC, CC and Fig+Tac = ~105.
So I guess I understand where you come from: In 84% you win and then +105 would be average then. Did I understand you correctly?

[JDOW]

@hobo:

I have NO idea where you get 58 or whatever.

Just to clarify: +58 (actually 59 after 10k simulations) is the average TUV-Swing the battle calculator spits out. I attached a screen.

the reason is the average defender hits in the first round have NO effect on the total attacker combat power in the second round since it’s just tipping ships and losing 0 combat attack value carriers.  since the average combat length is 3 rounds, what you choose as losses, ie keeping 1 carrier vs doing default losses, really has little effect on the total outcome.

Got you.

if you want to call this lucky go ahead

As I said, I do not call anything lucky. So far I just lived so far under the probably false assumption that an actual TUV score which is above the average TUV score by the calculator is a good result for the attacker. However I meanwhile have to acknowledge that in many battles an above average score is the most likely result.
Maybe also in this battle: You have 84% chance to win, if you win you have 7.14 units left which would equal e.g. 2 BB, 2 AC, CC and Fig+Tac = ~105.
So I guess I understand where you come from: In 84% you win and then +105 would be average then. Did I understand you correctly?

[JDOW]

I think this discussion helped me a lot to understand the outcome of battles.
I guess the reason of my false perception is that the median and the mean in such Monte Carlo simulations might be quite far away from each other.

[hobo]

yes you got it now, do not look at the average tuv swing.  there are 2 average battle scores, one when the attacker wins and one when the defender wins.  all you have to do to find the average battle score in each case is subtract the respective ipcs left from the starting value of the opposing side.  when the attacker wins the average battle score should be around 100 and when the defender wins it should be around 20(basically the defender would only have a tipped battleship left).  using default loses the attack should have about 8 hits left.  did you change the default ool to get the 7.14 or maybe that was calculated some other way I’m not sure.  anyways I’m sure you can see now that the results were about 2-3 hits off the mean, as I originally stated.

[JDOW]

I did not change the original OOL as you convinced me already that OOL does not have a super big difference on the overall outcome.

[gamerman01]

I just ran it by hand using my home-made spreadsheet, and I figure expected TUV difference to be about 110, which is almost exactly how it actually came out in the game.  So I agree with Hobo it went average.

[boldfresh]

the average result must take into account the fact the defender can win no?  so how can you say you were pretty much average given the fact that you were simply close to average presupposing an offensive win?

[boldfresh]

@Gamerman01:

I just ran it by hand using my home-made spreadsheet, and I figure expected TUV difference to be about 110, which is almost exactly how it actually came out in the game.  So I agree with Hobo it went average.

hobo said he came out 2-3 hits ahead so what gives?  are we saying that’s pretty much average?

[gamerman01]

I’m not pre-supposing an attacker win.  I just took expected hits and losses round by round by hand, and came up with pretty much the same survivors Hobo got, and a +110 TUV result.

[gamerman01]

@Boldfresh:

@Gamerman01:

I just ran it by hand using my home-made spreadsheet, and I figure expected TUV difference to be about 110, which is almost exactly how it actually came out in the game.  So I agree with Hobo it went average.

hobo said he came out 2-3 hits ahead so what gives?  are we saying that’s pretty much average?

Well, I saved planes in my simulation, and he lost them earlier.  I think that accounts for the couple extra hits.

[boldfresh]

i guess for a fleet battle (which so often skew crazy hard in my
experience) it is pretty close to average if it is within 2 or three hits. � :lol:

[gamerman01]

Point is, I agree that +107 is much closer to expected average than +58

[JDOW]

I think this comes from the fact the TUV change between “worst possible win for Axis” and closest loss for Axis is so big.
Worst possible win would be one BB remaining = +11. However once the defender wins it would be = -57.
This makes the median TUV-swing so much bigger than the average TUV-Swing and the Median-TUV Swing is closer to the most probably result.

So maybe the calculator would be better if it displayed the median.

And back to the original question: the battle went to odds then :-)