Flakking?

[Turnip2k]

Has anyone else thought about how effective flakking actually is in lethal fights? It crossed my mind that this may not be as simple as I initally thought it would behave - and it turns out to be pretty nice. If anyone else has done this sort of thing, gimme a shout - I would like a chat.

Turns out, that in any fight you can work out what ratios of units will maximise the damage done to your attacker. The maths is pretty standard (I'll pass it around if anyone wants a look), but the results are pretty cool. I was looking at very very simplified cases - it will be possible to do more, but I couldn't be bothered to plug all the variables and coupled equations into maple.

Ok, so look at the case of CW's and shields VS RPG's (you can take armour stripping into account, but that just makes it a little more complex). For an equal value of units going into the fight, what ratio of CW and shields will give you the most damage done? I took the ratio of RPG's : CW as 1:0.6.

For 1 million RPG's incoming, the best ratio is

Number CW = 2.633162432*10^5, Number Shields = 6.007825166*10^5

For twice your value incoming, the values are :

N CW = 7762.243543, N Shields = 6.032478491*10^5 --> Very low CW number due to them getting zeroed if you have anymore and not firing at all

(so you only have 500k rpg value to use)

For half your value of incoming ;

N Shields = 3.449572367*10^5, N CW = 8.770629486*10^5

(so you have 2m rpg value to use)


Etc.... Also, I did a quick run for AD / TL vs PA as another little try.

Equal value : N Dogs = 1.370858186*10^6, N TL = 1.044371272*10^6
2x Value : N Dogs = 1.273911915*10^6, N TL = 4.339132128*10^5
1/2 Value : N Dogs = 1.077235594*10^6, N TL = 2.338414661*10^6

So the general trend is - if you want to do the most damage and the majority of enemy fires before your major killers, change your flak numbers in response to the size of the incoming or target. If they are bigger than you, a higher proportion of flak in your mobs is needed - the smaller, the less flak you should have. Kinda expected, but nice to have some numbers on it!

May be a slightly random post, but just somthing I found interesting! One of those things I had to work out because it was annoying me... I'll probably do some more work on this in the future, but thats all for now!

p.s. Kill ratio of your unit in this is irrelevant I think, since your just trying to maximise the number of units that survive - but anyone who has good relative kill ratios for stuff, I can probably try to expand this method to a more expansive form.

 

[Polo]

Optimal LET flak is simple.

Buy pure Bikers.
Attack Poms without Hypnos.
Result: Optimal.

Gets you the most dangerous stat too.

 

[WildDisease]

Buy pure Bikers?
Gets you the most picked on stat too ? xD

 

[timthetyrant]

i bet scrappers would have a field day agianst pure bikes aswell

 

[cb1202]

Buy pure Bikers?
Gets you the most picked on stat too ? xD

and if ur in a good ally (which you should be if ur massing bikers) is also gets you the most helped out stat.

I honestly dont understand the point of this thread.

 

[Scorpio]

I think Turnip2k just wanted to get ppl to discuss their views & experience on LET flakking.

Imo, you don't need to have LET flak when in a good / active alliance. You defend against the routes you're good against (SO vs RPG, Robo vs SO/Thug etc etc) and the ppl will defend you with whatever you need if you got incoming. If you're in a low/inactive alliance then LET flakking probably is more interesting...

Orr if you're solo, but then you can get the flak killing issue. Ok that you have cr*ploads of shields and not so many CW, but then you're going to get WTFlakked...
It all depends on the situation, really

Interesting to see it in numbers though

 

[Ahead]

I honestly dont understand the point of this thread.

Well then you clearly don't understand the importance of LET flak.

Good post, and interesting stuff Turnip

 

[Martin]

I only use flak when it flaks the opposite style. E.g. use armour to flak health visa versa.
I see very little use for Shields or Attack Dogs (in their individual routes). Although Dogs are awesome last tick they die too easy for flak and it proves to be just as expensive.
Marines I only use to draw a bit of fire from Apaches to improve ratios when defending against RPGs and stuff.

I've never been a fan of flak and never found it beneficial for me to mass units for the sake of keeping other units alive (at the expensive of having less of the *dangerous* units).

 

[Turnip2k]

Well - this is kind of a situational answer to the age old question of 'What is the best ratio of units I can have?'. That depends on what your fighting. But in theory, if you tell me what you are fighting, it is possible to tell you what ratio of units to build in order to do the most damage.

I just need approximate kill ratios for each unit firing on each other - and this will be problematic for routes which are heavily depandant upon armour stripping (mil, SO, bunker, HV) - but for routes without very high AD, very low HD units (Terror, fantasy, puppets, robo) it should be possible to do this quite accurately. Even with armour strippers, its still possible, just a bit more complex.

I'll try and do some more this weekend. It's not a 'battle calc' par se, it won't tell you how the fight is going to go exactly. It just tells you what you need to do to hurt the other guy the most.

@ Martin - if your units fire well before the major enemy killers fire, then this is a safe bet. However, if your units fire after some of their killers (eg. your a robo hitting a terror), then there will be a balance between having strong killers, and having let flak which ensures that the maximal number of your killing units (weighted by how much damage they do) fire at the enemy.

 

[rooney]

I only use flak when it flaks the opposite style. E.g. use armour to flak health visa versa.
I see very little use for Shields or Attack Dogs (in their individual routes). Although Dogs are awesome last tick they die too easy for flak and it proves to be just as expensive.
Marines I only use to draw a bit of fire from Apaches to improve ratios when defending against RPGs and stuff.

I've never been a fan of flak and never found it beneficial for me to mass units for the sake of keeping other units alive (at the expensive of having less of the *dangerous* units).

i agree with martin here, i would rather have 10mil CW than 5mil CW and 10mil shield, but thats a personal preference i guess. marines however, i do use a fair amount, but i think thats different since they also can be classed as 'dangerous' units.

 

[cb1202]

I don't really like the idea of using armor as LET flak for armor. The only time I use LET flak is with humvee, jeeps, or marines. That is also only while playing as a solo or in an inactive alliance. In active alliances with lots of robos or strikers, buying LET flak isn't really a concern, as those robos will provide flak for your units.

 

[Turnip2k]

If its purely damage done to the enemy that you are concerned with (at least for one tick - another tick and you need to factor in killing the maximum amount of enemy units which do the most damage to you as well) then the only thing that matters is maximising the number of your units which do the most damage when firing at the enemy.

@ roony

For your CW / shield example, I did a quick calculation for hitting pure TL's. Since TL's fire first, shields will factor into this non-trivally.

For a fair fight (10m TL enemy score),

NCW0 = 3.625895*10^6, NS0 = 25587 --> This tells you that if you ever have a fight like this, and you have the bigger available mob score - just mass CW's. Dont bother with shields.

However, the moment you are trying to kill a larger score of units than the amount you have avilable, shields matter. If you have 90% of the attackers score, the best mix is :

NS = 3.238642718*10^5, NCW = 3.140237343*10^6

As you can see the number of shields preferable goes up rapidly. What is interesting, is that as you continue to go down in your score compared to theirs, the number of shields oscillates again.

At 80% score NS = -4.102564103*10^5, NCW = 3.076923077*10^6 --> A negative number of shields means that you cant have a positive number of them to maximise damage - so just mass CW.

70% NCW = 2.580773514*10^6, NS = -86335.25595 --> Shields going back up

60% NCW = 2.093491562*10^6, NS0 = 2.159095145*10^5

25% NS = 9.842818208*10^5, NCW = 5.064301642*10^5 --> Now, more shields than CW.

In the above, I assumed the CW killing TL ratio is 1:0.35 and the TL killing CW ratio is 1:0.1

Wouldnt mind a chat with anyone who is proficent in maple / interested in this - I'm getting some odd results with other calcs.

 

[lavadog]

I know "some" maple (Engineering student), so if you're having trouble with the program, maybe I can help

 

[willymchilybily]

2 things can u rearrange this information such that its in terms of ratios

so incoming is rpg. assume £10billion worth or something

you have 3 headings

enemy troops incoming score (5%-333%)
ratio of Cw: sheilds
cost of troops

so you have £10billion incoming rpgs the troops represent 20% of your score attacking upwards. (optimum cw;shield ratio) (cost of the troops 100%/20%*£10,000,000,000 = £50bill worth of cw;shield)

then £10bil inc. rpg. but this time the guys troops score is 100% of your troop score (aka you have £10bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/100%*£10,000,000,000)

then £10bil inc. rpg. but this time the guys troops score is 300% of your troop score (aka you have £3.3bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/300%*£10,000,000,000)

the second thing ill save for another time actually. keep that to myself for a bit longer.

 

[Martin]

The optimum ratio is as few LETs as you need to kill it in a tick, or with no more than 1 tick of losses, protected by as much flak as possible. That's not mathmatics, that's knowing the ratios units kill.

Nice maths skills though willy, but irrelevant?

 

[Turnip2k]

2 things can u rearrange this information such that its in terms of ratios

so incoming is rpg. assume £10billion worth or something

you have 3 headings

enemy troops incoming score (5%-333%)
ratio of Cw: sheilds
cost of troops

so you have £10billion incoming rpgs the troops represent 20% of your score attacking upwards. (optimum cw;shield ratio) (cost of the troops 100%/20%*£10,000,000,000 = £50bill worth of cw;shield)

then £10bil inc. rpg. but this time the guys troops score is 100% of your troop score (aka you have £10bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/100%*£10,000,000,000)

then £10bil inc. rpg. but this time the guys troops score is 300% of your troop score (aka you have £3.3bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/300%*£10,000,000,000)

the second thing ill save for another time actually. keep that to myself for a bit longer.

I'll get round to that at some point later - been a bit busy recently.

@ Martin, that only works if you can zero them in one tick, which in most cases you can't if its a reasonably even fight, or if you are being hit by a route that will kill off alot of your units before you even fire (hence the reason I chose RPG vs shields). As for ratios, yes those are still definately needed for these calculations, but theres other stuff to consider in there too.

 

[Martin]

2 things can u rearrange this information such that its in terms of ratios

so incoming is rpg. assume £10billion worth or something

you have 3 headings

enemy troops incoming score (5%-333%)
ratio of Cw: sheilds
cost of troops

so you have £10billion incoming rpgs the troops represent 20% of your score attacking upwards. (optimum cw;shield ratio) (cost of the troops 100%/20%*£10,000,000,000 = £50bill worth of cw;shield)

then £10bil inc. rpg. but this time the guys troops score is 100% of your troop score (aka you have £10bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/100%*£10,000,000,000)

then £10bil inc. rpg. but this time the guys troops score is 300% of your troop score (aka you have £3.3bill worth of cw;shiled) (optimum ratio) (cost of the troops 100%/300%*£10,000,000,000)

the second thing ill save for another time actually. keep that to myself for a bit longer.

I'll get round to that at some point later - been a bit busy recently.

@ Martin, that only works if you can zero them in one tick, which in most cases you can't if its a reasonably even fight, or if you are being hit by a route that will kill off alot of your units before you even fire (hence the reason I chose RPG vs shields). As for ratios, yes those are still definately needed for these calculations, but theres other stuff to consider in there too.

So therefore it's still not as clear cut, as every situation is different. It's like asking what kind of shoes do you wear when in a car race? It's errelevant as there are sooo many other variables.

 

[Turnip2k]

Not really? Example. You mass CW's against RPG's, and alot will die before they fire. You start mixing in shields, and you have less CW's to start with (assuming you have the same cash to spend), but more CW's survive to fire in alot of situations (see the numbers in previous post). Therefore you do more damage. Too many shields, and you don't have enough CW initally do get the damage done, even though a large fraction of what you bought survive. I'm trying to find the point at which you get just enough shields to allow the absoloute maximum number of CW to fire back.

Lethal flak lets your killers fire - finding the right amout of lethal flak to get the most killers to fire in each case is the point of the maths.

The only true unknown variable is the unit on unit kill ratio (and armour stripping etc...if i get round to it). Cost also factors into this too, for all units involved, as well as what fires before what, and what units have paricular resiliance to certain units (as you add more units firing).

 

[Martin]

Thing is, you could set up to fight 2mil RPGs and then 4mil RPGs come at you.... therefore you can get no set ratio....

 

[Turnip2k]

Nah, but if you have any money, it tells you at least what units to buy to inflict the most damage back (even if it isnt the optimum ratio, you can head that way).

Admittedly, CW / shields vs rpg is going to run into problems quickly, because CW will just get zeroed before they first on the first tick if you have any reasonable number of them (as the numbers before show).

I'm not dictating everyones exact ratio for every single incoming them could possibly get - but by giving them a reasonable trend in numbers, they know what to buy more of should they want to attack / defend against a certain incoming.