Experiment on naval combat results
I decided that, before messing with the PATE combat formulas, I'd better lay down a baseline for what the results are now, so I can tell how much I change them. Accordingly, I set up code so that I could run 48 identical battles at once, and did it twice so that I got 96 battles for each scenario. I used 4 different scenarios: 1 SL vs 1 SL, 6 SLs vs. 6 SLs, 6 SLs vs. 1 SL with the lone SL to windward, and 6 vs 1 with the lone SL to leeward. (I don't think using FRs instead would change anything.) In each scenario I did four variants: range/aim rigging, range/aim hull, close/aim rigging, and close/aim hull. In the 6 on 6 battle I also did broken lines/aim rigging and broken lines/aim hull. (Broken lines doesn't mean anything when one side has only one ship, so I didn't do those variants in the scenarios with one or both sides having 1 SL.)
I'll present the results separately for each scenario, but here I'll report one overreaching result: In 864 battles in which all ships on both sides fired at rigging, there was never, ever, a sunk ship. This seems incorrect. I think the reason is that total damage is maxed out at 10 or 11, maybe 12, so that when 1/3 of the damage is to the hull, the max that can ever be received is 3 or at most 4, so no ship ever sinks. I will probably want to change that somehow.
Three more notes will have the results for 1x1, 6x1, and 6x6 (I'll combine both variants of 6x1 into one post.)
Steve
Who's new
- joequ1oeil
- koolguy66
- Sakkal
- David Adams
- rkimler
Just tuned in...
Steve,
Thanks for doing all that work - everything that I see is well-described and pretty clear for those of us that have no idea how code works. I just got back onto the forum, so I'll peruse things more in depth and try to add relevant comments from the British experience in NWOL3, though you seem to have most of that at hand from Jim's copious remarks.
Sounds good
I'd like to hear from as many of the British players as care to comment (and non-British players too! But I know the British team thought about this a lot.) Be great if Jim V would reappear before this is all over with, but in his absence, definitely others should step in.
Steve
I am a wizard. I make things using magic. SJS
Pursuit
Steve,
One of the big bugbears in NWOL3 (and I'm sure this is on your to-do list) was naval pursuit. Do you have some thoughts on how we could get this to more accurately reflect what we (at least we ex-British!) think should be happening?
Pursuit
I looked at that briefly after NWOL 3 ended, and I am fairly sure that there is a design error. Consider the following case. FR1FR encounters BR1FR. FR1FR wishes to evade, BR1FR wishes to pursue. As FR1FR sails off with BR1FR in pursuit, they encounter FR2FR, also in evade mode. There is code to permit FR2FR to get away from BR1FR and FR1FR, as it should be - the two French ships just have to sail apart and BR1FR can only stick with one of them. I think the problem is that that code is allowing FR1FR to re-evade at that moment also, which it should not. All I have to do is make it so that the code that lets FR2FR evade knows to bypass FR1FR which is already "caught" by BR1FR. Should not be hard to do.
I need to have a more careful look at the code and run some tests, but changes are intended here.
There is also a bug with automatic detection by ships that don't move during a turn (Jim V tells a lovely saga of the four British frigates that spent the entire game in the Gulf of Lyon and never found anyone) that I need to fix as well.
Steve
I am a wizard. I make things using magic. SJS
Too many sinkings
Steve,
The numbers for sunk ships is too high in my opinion.
In most cases if as ship was nearing the point of sinking, it would surrender and be captured. Much in the same way that if your city walls were breached and were about to get massacred on land you would surrender. Sinking would generally occur only if the ships caught fire or exploded, which generally happened quickly and didn't give the captain the option of surrendering.
This also means that the ships that should be captured at the end of the battle are the ones that are badly damaged. It shouldn't be a random roll of all enemy ships. It is much like the saying you don't have to be faster than the bear, you just have to be faster than your friend. The damaged ships will always be captured before the undamaged ships.
capture
The capture chance is indeed dependent on the amount of damage the ship has taken. I will check to see if maybe it should be a little more dependent. I know it depends more on rigging damage than hull damage, which is proper.
Steve
I am a wizard. I make things using magic. SJS
Add surrender?
Can you put in some code where if you take 4 hull damage and you are on the losing side you the ship needs to pass a morale check to avoid striking the colors?
Plus, I'd make ships with 0-2 rigging damage nearly immune to capture. Historically most of the ships taken as prizes were not immediately pressed back into service by the new owners. Many of them we so badly damaged that they simply scuttled them rather than repair them (which historically was not free, like it is in NWOL). Capturing intact enemy ships also skews the game balance, since the new owner's fleet is now instantly more powerful than it was before.
-nick
Repair
Repair should definitely be costly. On my list of things to attend to at some point.
Steve
I am a wizard. I make things using magic. SJS
Experiment results: 6 on 1 battles
Here are the results for the 6 on 1 battles. (5 of the 6 are British in these runs, so the battle is a little more one-sided than the ship count would indicate; about 8 to 1.) I did a total of 16 runs, 48 battles per run, for a total of 768 battles; the 1-ship side won 2 of those 768. This percentage does not offend me. Sometimes the commander of the 6-ship squadron, or his captains, screw up royally or are cowardly, and the errors are compounded by wind shifts and a lucky shot into the magazine, and a miracle happens. As long as it happens one time in 400, it is not unrealistic.
First I'll present the results with the 1 ship is to leeward (the position from which it can escape most easily). Prizes is the total number of prizes taken (always the 1 ship; there was not a capture in the 2 battles the 1-ship side won, though conceivably it could happen), losesink and winssink are the number of ships on the losing and winning side sunk, respectively.
With the 1 ship to leeward:
Run Battle Aim Close Prizes Losesink Winsink
1st 6 v 1 Rig No 12 0 0
2nd 6 v 1 Rig No 5 0 0
1st 6 v 1 Hull No 9 7 2
2nd 6 v 1 Hull No 6 6 1
1st 6 v 1 Rig Yes 26 0 0
2nd 6 v 1 Rig Yes 15 0 0
1st 6 v 1 Hull Yes 17 ? ?
2nd 6 v 1 Hull Yes 17 11 4
(? indicates data I failed to record.) These all seem fairly reasonable. You catch him more often if you aim rigging (though maybe not as much more often as you should; same point as obtains in the 1 on 1 cases) and in the aiming at hull case you sink him reasonably often. If you get close your chance to catch him is about 1 in 3 if you aim hull, about 1 in 2 if you aim rigging. With the ship to leeward he should have good chances to escape if you don't close, and still reasonable chances to escape if you do.
With the 1 ship to windward:
Run Battle Aim Close Prizes Losesink Winsink
1st 6 v 1 Rig No 14 0 0
2nd 6 v 1 Rig No 13 0 0
1st 6 v 1 Hull No 4 9 6
2nd 6 v 1 Hull No 6 10 2
1st 6 v 1 Rig Yes 24 0 0
2nd 6 v 1 Rig Yes 27 0 0
1st 6 v 1 Hull Yes 20 15 4
2nd 6 v 1 Hull Yes 14 18 3
About the same pattern, although you sink him more often (which seems a little off to me, since the windward ship gets the strength bonus: but it's harder for him to run when he's upwind and maybe that gives you better odds to sink him?) He again takes one with him about 1 time in 10. When fighting close, the number of captures is about the same as it was when the 1-ship was to leeward, which is correct; once you reach close range he can break to the leeward side whether he had it at start or not. However, when you engage at range, the chance to catch him, although double what it was when he was to leeward, is still pretty low (and I think this was one of Jim Voege's points, though I think Jim thought the 1-ship should almost never get away, which I think is not the case.) This too suggests that I need to look at the escape chance formula.
However, I do -not- see any evidence that the side with the 6-to-1 advantage is losing more ships than it should. I didn't look at damage totals (much too hard to do) and maybe they took more damage than they should have - but it didn't result in a bunch of sunk ships. So maybe the problem here is not what I thought? Or maybe a 6-to-1 (8-to-1 in strength) is too large an advantage and if I did it with 2-to-1 or 3-to-1 it would materialize. British players, when you saw outnumbered ships sinking too many of yours, about what range did the odds usually fall into?
Steve
I am a wizard. I make things using magic. SJS
Experiment results: 1 on 1 battles
The single ship combat results are as follows (There are 48 ships total involved on each side in the 48 battles, prizes is total number of ships captured in the 48 battles, losesink and winssink are the total ships on each side sunk.)
When engaged at range:
Run Battle Aim Close Prizes Losesink Winsink
1st 1 v 1 Rig No 4 0 0
2nd 1 v 1 Rig No 5 0 0
1st 1 v 1 Hull No 5 6 2
2nd 1 v 1 Hull No 6 2 3
Most of the time nothing particular happens except that one ship leaves. It is alarming that when a ship did sink, it was the "winning" ship more than a third of the time. I think I need a tweak here, as noted in an earlier post. I also think there should be more captures when firing at rigging.
When engaging at close quarters:
1st 1 v 1 Rig Yes 17 0 0
2nd 1 v 1 Rig Yes 15 0 0
1st 1 v 1 Hull Yes 13 6 1
2nd 1 v 1 Hull Yes 21 4 1
Here it is much more likely that a ship is taken - it happens in almost, but not quite, half the battles - but not much more likely that one sinks. Again it seems like firing at rigging should be having more effect on the number of captures. For reasons not clear to me, the winning side had a lot fewer ships sunk in this case than in the range case. I think this is a random fluke, and the tweak should probably apply in both cases to make sure weird things don't occur.
Given the tweak to make sure the "winning" ship doesn't sink or at least does so quite rarely, I think these results are OK, but I do want to check the capture chance to make sure it's as sensitive to rigging damage as it should be, and probably tweak that too.
Steve
I am a wizard. I make things using magic. SJS
Experiment results: 6 on 6 battles
Note that the total ships involved in 48 battles of 6 on 6 is 288 on each side. At long range the results were as follows (Prizes is the total ships taken as prizes in the 48 battles, Losesink and Winsink is the total ships on the losing and winning side, respectively, sunk in the 48 battles.)
Aim Close Prizes Losesink Winsink
Rig No 47 0 0
Rig No 39 0 0
Hull No 18 31 23
Hull No 25 37 18
So at long range, you capture about 1 ship per battle when aiming at rigging, and nothing sinks. When aiming hull, you sink about 0.5 ships per battle, the loser loses more ships than the winner does in a ratio of maybe 3 to 2, but not very many ships are sunk, both sides total, just a little over 1 per battle. This all seems reasonable to me.
At close range line ahead:
Rig Yes 130 0 0
Rig Yes 125 0 0
Hull Yes 78 66 32
Hull Yes 83 86 43
Once you get to close range things get a lot bloodier. Firing at rigging, no one sinks, you capture about half the losing side ships, 3 of the 6. Firing at hull, you capture only about two, you sink about two, you lose one of your own on average.
If you break the lines:
Rig Break 128 0 0
Rig Break 130 0 0
Hull Break 76 112 68
Hull Break 72 112 58
Firing at rigging, it makes no difference whether you break lines or stay line ahead, either way you capture a bit under half the enemy ships. You do sink a lot more with broken lines, both sides lose about 50% more ships that way than they did in line ahead.
Except for not losing any ships at all when firing at rigging, this seems like a reasonable set of results to me. I am not sure exactly what difference I would expect between a battle in close quarters line ahead and a battle with broken lines, but more sinkers and the same number of captures doesn't strike me as obviously wrong.
Steve
I am a wizard. I make things using magic. SJS