Balancing Hi-Sec Freighter Bump Mechanics

Those Jita ganks are mostly concentrated in a certain event. And during that event the organizers have hundreds of ships ready. Hardly a counter-argument. And as long as there are mostly ships on a kill mail and no or only few pods, the gankers brought too many ships, which they only do to support arguments like your’s and to justify their whining about how hard ganking is.

If they use the wrong ships, that’s not a counter-argument either. They have the correct ships ready. If they don’t use them, that’s their fault and not the fault of the EHP of the freighter.

Nope. The data is the last 2000 ganks. Way more than just Burn Jita. The week of burn Jita I ignored, because that would have inflated the figures for Jita.

Happy to have you go look at the data and show it’s wrong.

Fault isn’t relevant. It just what is needed in some situations.

You can ignore it all you like, but in some situations, upwards of 20 gank ships is needed to gank a fully tanked freighter and there is no situation where 6 ships can gank a fully tanked freighter. That’s just the truth.

Considering that only ~1500 freighters and jump freigthers combined (most of them to wars even) have been killed in Jita (at least according to current zkill numbers), I wonder where you get the number 2000 from.

What’s needed are the correct ships for the correct circumstances. CFCODE deliberately use the wrong ships to inflate numbers as ganks of more competent gankers prove. That’s the truth.

It’s the last 2000 ganks in highsec. It would be impossible otherwise to know it’s 1.5% in Jita.

Surely that isn’t hard to understand.

In terms of the data, I continuously take the Redis feed from zkillboard. I use it for my own intel on a site I run locally. That way I can always analyse any route I am about to haul on, to see if there are active gank fleets and active suidice points. I’ve posted screenshots of it before:

So based on what I’m flying, I can set it to pull out the kills and show me whether any particular system is a potential problem. That just gives me a heap of data to analyse.

And I just have a view, that people who offer a good, safe service, should be able to capitalise on the effort they put in. Like this guy for example:

https://www.reddit.com/r/Eve/comments/9612nt/my_review_of_jareds_hauling_service/

Why, when he is offering a good service, should someone else have a more even field in competing with him and be able to be lazy in how they haul. The less even the playing field, the better. Then we all live and die based on how we play the game and not how much the mechanics protect us.

Fair enough. Considering it’s you, anything is possible. And considering that you include CODE numbers, you rely on inflated numbers to support your argument. Killmails show that no 20 chars are necessary to gank a freighter in the main ganking systems and CFCODE only uses more to inflate these numbers to support their claims on how difficult ganking is.

As said in the other thread, they are not even consistent with that as they can gank the same kind of freighter with the same kind of fitting with 50% less people depending on what they feel like. They even use fewer people on higher value targets than on lower value targets. Just as an illustration what kind of bogus numbers you base your argument on:

This guy is doing frigate size and BR size contracts. What’s your point?

That people who are lazy don’t deserve more assistance for being lazy. We should be left to sort out the problems we face in the game and in general, proposals we make as players for change should both give and take, not just give to one side and take from the other.

Gankers are lazy by relying on non-concordable tackle and residing in unassailable NPC stations. Your point is? If anything, I would say, both sides are guilty of laziness. Can we agree to that?

That people who are lazy don’t deserve more assistance for being lazy. We should be left to sort out the problems we face in the game and in general, proposals we make as players for change should both give and take, not just give to one side and take from the other.

That statement doesn’t favour one side over the other. Make the game more challenging for us all. That would be great.

1 Like

wtf? That 11 ships is the most common? It’s the modal value (ie. the most common. Not the mean or median).

Ah yeah, consdering it’s you, I should have assumed you didn’t know what it meant. My bad.

It is not about most common, it is about what is possible. And the numbers you base your argument on are blown way out of proportion compared to what is possible. This is evident by the above 2 examples, whose pattern is pervasive thorough the freighter gank killboards. Your numbers are simply inflated and therefore not credible.

The numbers that there are 1.5% of ganks in Jita.

You are trying to argue a point I’m not really making. Sure, in some cases fewer ships are needed. In some cases, more are needed. But 1.5% of ganks occur in Jita, and in that system there are absolutely situations where more than 20 ships are needed in main gank systems (even in Uedama in some situations).

So off 30 ganks where you have no idea of the skills of each pilot or how much overkill they had or if they threw double the pilots at it because there were some anti gankers sitting around the station also trying to stop them, you are judging how many pilots are ‘needed’. Without any kind of maths based evidence.
Really…

Wrong.

If you take all level 5 skills (which is the normal in these types of calculations) and apply the same to a Freighter fully tanked, there are many situations where more than 20 ships are needed.

I think you are trying to argue I’m wrong, without even doing any math yourself.

So if you go run a range of scenarios (eg. catalysts to gank a fully tanked Obelisk in systems from 0.5 - 1.0, all of them), you’ll see there are more than 20 catalysts needed.

Go do the math and I’m sure you’ll see the same thing. There are also scenarios where fewer are needed. It’s not an absolute and I’ve never claimed it is.

No, that is not what I am arguing. I am arguing that CODE uses vastly different numbers of gank ships to gank the exact same target with the exact same ship fitting in the exact same sec status system in the exact same scenarios. There are no different circumstances that support your claims. The numbers are just intentionally inflated.

Another example:
Providence | Peter Ramen | Killmail | zKillboard - untanked, 30+ chars in cheap ships
Providence | Michael D-Delta | Killmail | zKillboard - fully tanked, 30+ chars in cheap ships

Another example:
Providence | Vad78RUS | Killmail | zKillboard - istabs, 30+ chars in cheap ships
Providence | Hauler Jill | Killmail | zKillboard - untanked, 30+ chars in cheap ships

And I can go on and on and on, 2 things will stay the same: The numbers you rely on are complete bogus. You do not need massive numbers of chars to gank a freighter.

1 Like

I wasn’t aware catalysts were the only ship you could gank in. Could you show me the update where CCP changed the rules on ganking?

Forget CODE. Forget any other gankers. Just go off the calculations and different scenarios.

CODE aren’t relevant and the calculations are all based on level 5 skills across the board. No alliance or corporation changes that.

Why? If I do that, the outlook for your argument is even worse. Let me show you:
Providence | morria swista | Killmail | zKillboard - ganked with 10 guys and it was tanked.
Providence | KepHec QBit | Killmail | zKillboard - untanked and ganked by 6 guys
Obelisk | Buliwyf Khan | Killmail | zKillboard - untanked and ganked by 5 guys

Or let’s go back to CODE in a higher security system:
Charon | Habdich Lulu | Killmail | zKillboard - untanked and ganked by 27 Bombers
Charon | Lysithea Emit | Killmail | zKillboard - slightly tanked but ganked by 69 people in Bombers and Talos, and it was cheaper than the first freighter.

The numbers you rely on are bogus and CFCODE intentionally and deliberately inflate them to support the argument how hard ganking is. The killboard proves the opposite.

If you take my own Freighter pilot for example, with me webbing (here’s my normal web fit minus the armour command bursts I also normally use):

There are a tonne of situations where it takes more than 20 gank ships to kill her if I was ever unlucky enough to end up in that situation (I rarely need to fly the Obelisk and mainly stick with my Anshar which is near 1 million EHP).

It doesn’t matter who tries to gank it. The situation is the same.

No inflation of numbers for CODE or anything. Just all based on the EHP my freighter character can achieve in her freighter.

Now, I have never claimed it only ever takes 20 ships to gank a freighter in all cases, but somehow you seem to have read the words “some situations” as all, which isn’t what was written. Not even close.

Your armour boosts are on that.

Of course they are and they generally always are. Plus the high-grade slaves. I use the mechanics to the best advantage I can. So do others. I also use the Skirmish burst where needed to get overheated 57km web range, while running the armour burst when I need and with a sub 1-second lock time and 3 ticks into warp.

You can try to artificially limit what I wrote, when I wrote there are some situations where it takes upwards of 20 ships to gank a fully tanked freighter, all you like. But the game doesn’t care and just as bumpers and gankers don’t play fair, we don’t have to either.

So if I use links, implants and even boosters if I need…then good on me in that case. That’s best use of the mechanics.