Battle calculator issue
-
Not sure if its and issue or not but i don't understand the results of the battle calculator on a sea battle in which transports are also involved. AFAIK transport cant either attack or chosen as causalities.
So first question is when using the battle calculator and in case a transport exists on the SZ the battle is conducted shall i remove it in order to get accurate results? Additionally pls check the attached screenshots. i am getting different results by including or not a transport on calculation.
By adding a transport on defender, the attacker has e.g. 90% chance to win. When i remove only the transport the chances of attacker dropped to around 50%. Is that normal or its a known bug and should be reported?
Thanks in advance.
-
@fatoumas even though they have hitpoints the simulator will not take them as casualties unless the map allows it. you will notice on a map where they are only allowed as last casualties that it will not make a difference whether you have in there or not.
-
@ubernaut Thanks for your fat response. ok. So on that map are not chosen as causalities so i will consider that they dont count in the battle. But what about the calculation? what is the correct one? and why the attacker has more chances to win if the defender adds 1 transport?
-
@fatoumas try it and see it will not affect the outcomes.
-
@fatoumas said in Battle calculator issue:
Not sure if its and issue or not but i don't understand the results of the battle calculator on a sea battle in which transports are also involved. AFAIK transport cant either attack or chosen as causalities.
So first question is when using the battle calculator and in case a transport exists on the SZ the battle is conducted shall i remove it in order to get accurate results? Additionally pls check the attached screenshots. i am getting different results by including or not a transport on calculation.
By adding a transport on defender, the attacker has e.g. 90% chance to win. When i remove only the transport the chances of attacker dropped to around 50%. Is that normal or its a known bug and should be reported?
Thanks in advance.
Moved to "Feature Requests & Ideas" because I believe the matter does exist, so this would be a feature request for a better battle-simulation accounting.
@fatoumas First of all, let me introduce you to the fact that TripleA absurdly (nothing personal against whoever decided that 200 was good, but 200 is like on purpose to make the battle-calculator work badly) has 200 run count default. In dice, that can easily cause wildly different results when calculating odds.
In "Engine Preferences", I suggest you change the default at least to 1000 (1,000), even though myself I use 10000 (10,000) because also 1,000 is not really good enough: you can still get several percentage points more than average once in a while (like getting a 48% instead of a 52% sometimes happens at 1,000 run count).
By the way, here is the issue where I already said that the default for dice run counts should be increased:
https://forums.triplea-game.org/topic/863/increasing-regular-dice-default-run-count
This said, I remade those battle-simulations myself, but using 10,000 run count instead of 200, obtaining 55% (instead of your 52%) in the first case and 82% (instead of your 86%) in the second case (By the way, my simulation time has been 2.47 seconds only at 10,000 simulation count.).
In both cases, however, the probability of the defender to win is 6% (as you also obtained in the first case, whereas the 2% you obtained in the second case has to be a wild shot you can easily get using a 200 run count).
My best guess is that TripleA (of course) counts the battle as a draw if ending with only air units against only submarines (this is indeed a stalemate) yet it counts it as a win if at this point also one or more transports remain and are automatically destroyed. I agree that it would be good if a developer can make any sense of this, as I agree adding transports to the defender ought not to increase the winning chances of the attacker. In practice, it should give you 55% in either case, yet the matter here would be differentiating between "draw" as in when every side has 0 units surviving and "stalemate" as in when every side has 1 or more units surviving none of which can hit any enemy units in the battle (which is air-only against submarines-only). In my opininion, adding a defending transport in your situation should actually increase the chance of the defender to win, as it should allow for a small chance of zero offenders remaining and the transport surviving as the only defender. If TripleA is actually accounting such an outcome as a draw too, this should be clarified (maybe as a tooltip for the "Draw", telling exactly what "draw" actually means).
What I said I think is confirmed by adding more transports to the defender.
- With 0 transport, the win/draw/lose outcomes are 55/39/6%.
- With 1 transport, the win/draw/lose outcomes are 82/12/6%.
- With 100 transports, the win/draw/lose outcomes are 94/0/6%.
I believe this shows that the air vs. submarines stalemates is accounted as a win for the offender if, at the point in which that happens, there is still in the battle a defending transport automatically to kill. If my interpretation is correct, it would be, of course, nonsensical, as it makes no sense that the battle is counted as a draw if the transport is already killed before that point, whereas it is counted as a win for the offender if the transport survives to be killed at that point (the actual outcomes are exactly the same). The fact that it gives a 0% draw with 100 transports makes me guess that the outcome of all combat units being destroyed while all or some of the transports survive is counted not as a draw. However, I'm not sure how it is counted instead. If it is actually counted as a win for the offender, I would say this would make really no sense: it should be counted either as a draw or as a win for the defender, and the program should inform the user on how transports-only surviving is actually accounted.
-
@cernel Thanks for the detailed reply! One more question pls. What the difference between the "ave. attacker units left" vs "Units left if Attacker won". Is the first one the average while the other one the best case scenario how many units will survive?
-
@fatoumas the first averages all battles, the latter averages only the winning battles.
-
@ubernaut ty!
-
@fatoumas np
