Enemy bot targeting question

[Musket22]

I've been wondering if anyone has any insight about how bots prioritize targeting enemy ships.

This AM, I'm in a Sov CA (1/2 health) with 2 friendlies (BB & CA) neither at full health about 3-4km ahead.

Beyond them about 4km away is an enemy-bot BB.

This is all open water for 10km, so all ships visible. 

So with me being the farthest, why am I the bot's target of choice?

Edited March 2 by Musket22

[Wrath_of_Deadguy]

Range to target is a much lower priority in their calculations than HP is. If you had less HP remaining than your teammates, even if they had taken damage too, that was probably the reason the bot focused you instead of them.

[SunkCostFallacy]

I've played a ridiculous number of games in COOP, and have been pretty motivated to figure out this very question because it made little sense to me.

I'll offer the deductions that I've made so far, with the proviso that this is just what I think, with no real evidence other than my experience in-game. On top of that I run Priority Target on almost all of my cruisers (and I'm a cruiser main) so I know how many bots are targeting me, and I pay attention to that number so I know when it changes and take note of what I did that caused it to change.

1. Bots will target the first available target after a match begins.

In the feeding frenzy of COOP I frequently charge forward in my Salem (which has a lighthouse build) and as a result it's not uncommon that, the moment I'm spotted, I will have five bots targeting me. As soon as they get the chance, every bot who can will target the first available target unless another target has appeared.

2. Bots will continue to aim at their target until they can't.

As per the above example ... I will continue to be the target until a bot loses LOS or spotting. I've proved this in multiple tests by charging forward, getting focus fired, then breaking all LOS by ducking behind an island. As long as you break LOS/spotting for long enough for Priority Target to show zero, when you reappear (which can be a matter of seconds) those bots will have switched to another target and they will NOT re-target you until something causes them to. Even if their guns are still pointed at you!

3. Given the choice of multiple targets of equal validity (whatever makes a target valid according to the algorithm) the bots will select the target with the lowest PERCENTAGE of HP.

To clarify this ... if a bot has two possible targets, a DD with 100% health, for example 12800HP, and a Cruiser with significant damage which has taken it down to 12800HP (i.e. the same HP as the DD) ... the bots will target the cruiser. This is harder to 'prove' because who has the chance to compare HPs during a match but it's demonstrable if you watch. Bots will invariably target the most damaged ship if all other factors are equal.

4. Distance is not a major factor, and potential threat is completely ignored.

I have received fire from ships that I have not fired upon, which are being pounded by ships which are closer. I can't even begin to count the number of times I've seen a bot fire over the heads of the ships which are attacking it, so that it can try to kill a ship which has a lower XP percentage, or because it was the first ship that bot was able to target. This is particularly noticeable with DDs. Bots will continue to target a distant target even when a DD is steaming up to unload a full spread of torpedoes into the bots broadside. At most the bots secondaries will work on the DD, but they won't FOCUS the DD because Bots don't use their secondaries that way.

5. The Bot programming is inconsistent.

Let me be clear about this ... I am not commenting on the quality of the code. What I'm saying is that in any given COOP match, there is a chance that one or more bots will behave as if they have different programming to the rest of the bots. There are times when a bot seems to actually "level up" and starts getting more accurate with it's shooting, and with it's 'tactical decisions'. But that's just an impression and it's hard to substantiate. That having been said, the 'rules' 1 through 4 seem to be pretty consistent ... but it's possible that there are occasional bots doing different things that go unnoticed.

6. Bot CVs know what the real AA rating of every ship in the match is and target based on it.

Bot CVs know if you have Defensive AA fire, and will chose an equal target which does NOT have it if one is available. I'm just adding this for completeness.

7. Bot Subs will target and ping you even if they have no intention of firing on you.

I've been in multiple matches where a bot sub is moving from the flank to the middle of the map, and is pinging me (therefore is targeting me and shows up under Priority Target) but is broadside to me and therefor cannot fire it's torps at me. Again - I'm just adding this for completeness. Let's be honest ... some Bot Subs are trolls!

I must stress ... I have no hard evidence to support any of the above. WG don't reveal their bots algorithms (nor should they) so we can only go with 'best guess'. But I've got over 18 thousand matches in COOP and I feel confident in my observations.

[Efros]

On point 5, I suspect that WG throws in the occasional 'super' bot just to keep us interested. A bot that is dialed up, the one that's going to hit you on its first volley and will consistently hit you every volley after that, or if it's a BB just wipe your health out in two shots from 18k+. Occasionally you'll get one on the green team making you feel completely useless as you watch it mop up 5 kills.

[Amaruk]

11 hours ago, Efros said:

On point 5, I suspect that WG throws in the occasional 'super' bot just to keep us interested. A bot that is dialed up, the one that's going to hit you on its first volley and will consistently hit you every volley after that, or if it's a BB just wipe your health out in two shots from 18k+. Occasionally you'll get one on the green team making you feel completely useless as you watch it mop up 5 kills.

I think there are at least 3 levels of bot 'intelligence' in Coop, perhaps more.

After almost 11K games during the last 9 years or so, my observations pretty much mirror those of @SunkCostFallacy