Naval battle bug when superior?

[Gullet]

I am Germany, first turn of game. WA has fleets in Eastern Med
(2 HF, 2 LF, 1 TR) and GE has based in Southern Italy (3 HF, 3 LF, 4 TR, 2 SUB). If I attack with ALL Germany's fleets (even tried incl. TR, but do not know if that should actually make a diff., either way same result with all incl. TR) the defending WA is ALWAYS allowed to take more shots at GE than GE is at WA. How can this actually be realistic?!

I started noticing this phenomenon when attacking the "kicked out" Dutch LF in North Sea. If you move in just the one German HF the ratio is 0,03/1,26 and 84% probability of occupying the zone, if you add on the 2 LFs it is 0,76/0,76 and 89%...seriously, come on!

In the East Med battle it has also happened that my HFs did not get to fire at all! (do not recall if this also was true for WA HFs). How can this be as HFs are allowed to fire at ranges 1-3, i.e. should technically be included at either chosen fighting range, right? As combat calculations now go, it seems as attacker's numbers do not really count, which they should.

/G

[WanderingHead]

There is the confusing feature that not all of the fleet actually takes part in combat. Rulebook page 72 explains how many fleets engage, a random quantity. This may explain HF not firing.

This confusing feature may be the explanation for the weird thing I've noticed which is that when I stack in HFs, and then later add a LF, I actually see my expected casualties go up, expected kills go down, and probability of success go down.

Test it out. Fire up the game for 1940. Take 1,2,3 HF from Italy to Eastern Med, expected kills go 1.16, 1.16, 1.16. No increase for adding HFs, although probability of success does go up.

OK, 3 HF yields 1.16 expected kills, 46% probability of success.

Not good enough for me, commit more resources. Add a LF ... expected kills is now 0.66, probability of success is 0%. Less than without the LF.

Something about this situation seems a little ridiculous, doesn't it?

[Gullet]

Bang! right on target! [:)]

I agree to the fact that adding an LF could actually increase probability of own casualties, taking into account that more of the defenders ships, i.e. LFs, will be probable to conduct combat (if I interpret rules correctly). However, what I do not agree with is that superiority in numbers is eliminated due to some "fairness" tweak in the combat calculator. It is not as the ships are taking a ticket and attacking/being attacked one at a time (which the manual does not imply either).

Even with the explanation with the N-die followed by the X-die in the manual I have a hard time agreeing that the defender gets to fire MORE times than a numerically superior attacking force. Even if a subamount of ships is chosen each side should at least be allowed to fire the same amount of times. Having said that I still believe that the calculations seem to favor the inferior defender.

Two observations after reading manual again:

1. Any present SUBs do not seem to count in the N-die roll, only HF and LF, why not?

2. The expected value for number of engaging units is equal to 25% of the larger fleet (50% x 50%), which leaves a numerically superior fleet at no statistical advantage as long as it is not more than 4 times greater than the defender (1/25%), realistic?

Examplified with the E Med battle (counting subs for ease of description):

The N-die
GE has more units, i.e. 8 vs WA's 4, therefore the N-die is 0 to 8. The statistically expected roll will be 4 (as zero seems to be included according to the text, which actually makes it a N+1 die to be correct).

The X-die
The statistically expected roll from the X-die (1 through 4) is 2.5 and hence, the expected number of engaging vessels is 2 (or 3), i.e. 25% (37.5%) of the larger fleet.

[Joel Billings]

There are a few things going on here. First, once it is determined how many ships will fight in a surface battle, that number is used for both sides. So if 8 are attacking 2, and 2 is the number allowed, both sides will use 2. If 3 is allowed, the larger side will use 3 and the smaller side will use 2 (as that's all they have).

Surface combat (as opposed to aircraft vs ships) was extremely rare in WWII. Even when large numbers of fleets are roaming areas, it is not that hard to avoid combat if you want to. The larger number of fleets should give you a better chance of taking control of the area, but it only gives you a somewhat larger chance of bringing the enemy into a battle where you can significantly outnumber him. That's why we have this mechanism.

We also have a rule that fleets should match up when possible HF vs HF and LF vs LF. This is because commanders will attempt to stay out of battles where they are at a disadvantage if possible. Remember we are dealing with vast areas of water and 3 month turns, so there is a lot of abstraction built into the system. If you are facing a Light Fleet all alone, it is better to just send in the Heavy Fleet given the system, otherwise you give the enemy something it can match up with. Remember, 3 month turns and these fleets aren't all sailing together around the area for the entire 3 months. The more you understand the dynamics going on, the better you can suit your strategy to the situation.

As for the combat analyzer. We have found that there are some strange results that can come out of it regarding win/loss percentage especially. We are working on those that we know of and hope to have it working better in the next patch. For example, if you send a plane in versus one ship, it will show a chance of winning control of the area, even though there is no chance you will actually win since aircraft won't stay to control the area anyway. So we know this is something we need to fix. This will also improve the AI play as the AI make use of the analyzer even more than most human players, so as we improve it's accuracy, we will be improving the AI. The analyzer works ok in most cases, but it does have a few glitches.

If you have an analyzer situation that makes no sense to you and you think it might be a bug, please email a save with it to 2by3@2by3games.com and we will take a look. We do want to improve it as best we can. Thanks.

[Gullet]

A detailed answer as always, Mr Billings, which is much appreciated.

A few thoughts:

I kind of guessed the thoughts behind the "reduction/randomizer in the number of engaging ships" scheme and I agree with the general idea behind it. However, it is not logical that as long as the picked number is less/equal to defenders numbers that both sides actually bring the same number of ships to battle. If 8 attack 6, there should be a possibility that, e.g. 6 ships are up against 3, or even 2 against 4 (alhtought probability should be very low i guess). As the function is now it is tweaked very much in favor of defender. Why should the defender be better at keeping his outfit together? Given vast areas, both attacker and defender could be split-up although the numerically superior should have higher probability of meeting enemy at an advantage.

If memory serves me correct, both Bismarck and Graf Spee were eventually hunted down by superior forces actively hunting them. Graf Spee's captain "incorrectly" sinking the ship off Montevideo because he thought he was up to a superior force, which at the time actually was inferior to the pocket BB (a few destroyers?), but more vessels were due to arrive shortly. This kind of proves two points:

1. Superior forces could hunt down inferior ones and swing their weight (although agreed, not very frequent)
2. A massive hunt could actually render a situation where the superior force was represented by a lesser outfit at first contact with the enemy (Graf Spee in Montevideo).

The easiest fix if you agree with my reasoning is probably to have a "weighted" die, where there is bigger weight attached to larger numbers (or more probable match-ups. i.e. maybe 8 should be less probable than 6 in an 8 vs 6 attack) and hence lesser to the smaller; by that the probability for the superior side being able to use the extra weight would be increased (I think the statistical ratio of >4-1 is way too big).

More appealing though (I think?) would be to have all "set-ups" possible, i.e. in an 8 vs 6, both 8-1 and 4-6 would be possible, althought low probability of extremes of course. On the other hand just matching up ALL ships against each other and letting random decide if they hit or not could actually solve it the best. I mean you always stand the chance of rolling snake-eyes whether you are superior or inferior and as the defender always has a stable defence value and a two hit system a lot of the slaughter-effect is already taken care of. I guess that would kind of make the current algorithm useless, but sometimes an easy one is more accurate than a complicated one...

Question unanswered from last posting:

- do subs count in the match-up? if no, why not?

/G

[Uncle_Joe]

Allowing that many ships to fire would seriously effect the Pacific balance of power. As it is now, the thing that keeps surface ships in check is the small chance of a decisive engagement. Is it worth spending all the supplies to commit a huge battlefleet only to see just a small battle? That is the risk you take by commiting surface ships rather than aircraft.

With increased chances of firing opportunities, large gangs of surface ships would overwhelm carrier groups that would, in reality, have just evaded them. The game's scale precludes that type of detail, but the abstraction works out really well.

Another thing that might occur from allowing more firing opportunities is the naval version of the '500 pound gorilla'. Mass up 3-4 BBs and 3-4 CAs and you have basically an unkillable force that would be suicide to attack without equal numbers. There is no way to attrite the force without losing a huge number of ships yourself.

Also, consider some of the points from naval history. Yes, the Brits DID bring superior firepower to bear on the Bismarck and it told. But, consider that if not for the one lucky torpedo hit, the Bismarck would have eluded nearly the entire British Home Fleet and attachments. In game terms, the Heavy Fleet would have sailed out, damage/destroyed 1 British Heavy Fleet (Hood/PoW) and then retreated after the Brit shot missed. What happened 'historically' was that the Brit Carrier Air reduced the defense of of the German Heavy Fleet, which was enough for the the one shot to hit.

Personally, I think the naval combat works out quite well. You cant count on ships to always decisively engage, but the 'Naval Points' system ensures that if you send enough into a Zone, you will take it away from the enemy. That seems far better than increasing the chances of slaughter just by piling more ships in.

[Gullet]

You have some valid points and maybe it was too fast a thought to propose to "put 'em all up" á la Axis & Allies.

OK, having played the E Med assault a dussin more times now, on very few instances is the DE player allowed to use his larger numbers (as would be expected), so far so "good", but on one occassion 3 HF, 2 SUB were engaging 2 HF, 2 LF. Two observations: the third HF went for the LF instead of finishing off the 2 HF and the SUBs went for the LF when SUDDENLY the WA TRPT fires back at a less defensive SUB (defending at 3, because already attacked once). Questions:

1. Why did not the residual DE HF pair off with one of the WA HF (would at least be preferable from an attackers point of view and would be coherent with the like-for-like approach)

2. Which ships were actually selected for battle? One interpretation is that 3 from each side (as SUBs seem to be "bonus") meaning 3 HF vs 2 HF + LF. Since both sides actually have more than 3 available, how come the LF which the third HF fired upon did not fire back at it? And if it was 3 HF vs 2 HF + TRPT, how come the HF did not fire upon TRPT? Is it that it actually was 3 HF vs 2 HF + LF, but that the LF is programmed to go for attacking SUBs if present? in that case, how come the god damn TRPT gets to fire.

3. If now one of the two WA LF was engaged in defending itself against HF would it not be more rational for the SUB to go directly for the TRPT - I mean that is actually the prime target, no? The LF are just a type of shield between the TRPT and SUB, and when the shield is tied up in defending itself against HF, why not sneak by and attack TRPT (the LF anyways chose to go after the SUB in its attack sequence).

4. Is there the possibility that zero ships are chosen when the N-die is rolled? I have not seen it yet, so I am just curios to whether it was a possible outcome.

For reference I calculated that, given a no-show possibility, that there is a 7.6% likelihood of DE attacking in greater numbers in the E Med as there is only the two possibilities of that (5 vs 4 or 6 vs 4). As I have pointed out earlier I believe this is too low, and I would gladly see a change where NOT ALL fleets shoot out, but that the probability for the superior side is upped and also allow for e.g. 3 vs 1 in the above case.

/G

[WanderingHead]

Joel,

I understand the realities the abstraction are trying to represent. Actually, when I read the manual I thought it made sense.

Now that I play it I am not so sure.

Mainly what gets me is this: how can it make sense that I can throw another ship into the engagement (a LF after committing 3 HF) and it actually lowers my expected kills?

That one fundamental thing does not make sense.

As an aside, it seems as though if you are confident that you can survive the return fire, then one way to get all your ships to fire is to throw them in one at a time.

[Gullet]

ORIGINAL: WanderingHead

Mainly what gets me is this: how can it make sense that I can throw another ship into the engagement (a LF after committing 3 HF) and it actually lowers my expected kills?

Is it not probaly due to the fact touched upon in Billings answer above, i.e. that there suddenly is a possibility that one of the chosen attacking vessels is an LF? In your first posting you mentioned the probability going down to zero in the E Med attack - I get 49% when having 3 HF + 1 LF (the lowering of hits is there though)

As an aside, it seems as though if you are confident that you can survive the return fire, then one way to get all your ships to fire is to throw them in one at a time.

I guess from the description we so far received that you would actually stand a risk of facing, e.g. 1 vs 4 ships in the E Med attack (although evidently a small risk). On the other hand, say you do attack 1 by 1, and on the first attack, LF vs the bunch, the WA fleet all attack, do the only get to receive fire in the oncoming bombardments (they have already attacked in first attack)?

I just tested a dussin times attacking E Med with 1 LF and in 4 instances did the defender get to fire upon it with more than one unit (and then only with 2). The expected number of times would be 4.6 (12 x 38.33%), so it seems reasonable althought the absence of 3/4 vs 1 set-ups makes me a little suspicious. One odd thing though was one attack where the LF was fired upon by HF, but returning fire on LF - an LF which never fired upon the German LF...


I'd say something is not right with the naval combat thingy [;)]

[Uncle_Joe]

It depends on what you want to do. If you want to inflict a lot of damage, but take a lot in return, then yeah, throwing your ships out one at a time is the way to go. If you want to take the Sea Zone, the more the merrier. Even if your extra ships dont fire, they still add 'Naval Points' for determining who will have to retreat.

To me, that is the crux. Adding more ships doesnt necessarily make your fleet more lethal, but definately makes them more likely to take control of the area. If you dont care who ends up in ultimate control, then send the fleet in smaller groupings.

I can rationalize this as an abstraction of smaller fleets outmaneuvering larger, more ponderous ones.

Also keep in mind that it was mentioned that the Combat Calculator is not always displaying 100% accurate information. There appear to be a few kinks in it that are being worked on for the next patch.

[WanderingHead]

ORIGINAL: Uncle_Joe
To me, that is the crux. Adding more ships doesnt necessarily make your fleet more lethal, but definately makes them more likely to take control of the area. If you dont care who ends up in ultimate control, then send the fleet in smaller groupings.

I don't know that adding more up front is definitely better than the piecemeal approach.

One at a time, throwing in BBs first, you clear out units that are damaged. I think that damaged units still count as present for retreat purposes (is that right?), so by clearing them out from every 1-on-1 battle you actually do a lot better. Then in the end you send in the LF mosquitos.

At any rate, that's sort of how it worked out when I played it.