EVE General Discussion

People keep throwing the term "risk/reward" or "risk:reward" around as if each area of space or activity has an actual "risk/reward" score. Since it makes sense for such a score to exist, and reward is easy to define (isk/hr), we should endeavor to actually define it.

The only issue is... defining "risk" as a number. What would you use as an objective heuristic for risk measurement in Eve?

Apparently "EVE is real" so no, now get back to work.

There is a risk figure in game, called "system securtiy". Starts 1.0 and ends at -1.0

No problem, you are welcome.

I's say the same thing: ISK per hour, just as a loss. Everything that is risk can be "bought back" with ISK. Ships, modules, clones, implants. Even sec status can be referenced to ISK. Just take the time it takes to recover the sec status and multiply by the ISK per hour you would have been making if you were not out working on your sec status.

The difficulty here is loss is much more unpredictable. You go for months in null, then lose your entire home when your alliance collapses. Traders have a sudden market upset. This makes figuring out the potential ISK per hour of loss quite difficult.

Risk = cost × probability.

Can you expand on the probability part? No trolling, no contention, I'm actually interested in an attempt to assign mathematical properties to the whole "risk vs reward" argument.

And is that equation the same for reward?....

The problem with complex probabilities is that hey can be hedged so we go straight into actual-banking business theory on that one; the simple ones don't really allow for that, but on the other hand, they're easy (if you have the data).

Since you measure reward in ISK/h, we should aim at having the risk on the same scale, meaning our probability will be measured in losses per hour. If you want to calculate, say, the risk for a Hulk miner being suicide ganked, calculate the number of Hulk-hours spent across all of highsec in a day and the number of Hulk losses across all of highsec in a day. Divide the latter with the former and then divide by 24. We now have our per-hour probability. Then multiply with the cost of a Hulk to get our total risk.

e: clarifications.

The thing is it entirely depends on your competence. Missioning in lo sec or npc null for example holds zero risk for people who aren't idiots. Mining in hi sec holds risk for people who are idiots. Really rather than risk what people mean is "how much do i need to pay attention".

I think I know... but it'd involve lengthy research and data.

You could, for the sake of time and efficiency in analysis and reporting, select 4 systems known in highsec, lowsec, and nul (separately, so your total is 12 systems for the study, 4 systems to a cohort) for heavy mining (i.e. the most miners ACTIVELY in belts).

Once you have your systems selected, you'd need to (by some miracle) find a 100% accurate killboard or database of ships (namely, ship types) attacked and destroyed in system. Attacked versus destroyed is important here, as it defined successful ganks versus fail ganks. Ship types would be limited to mining/industrial ships (orca, iteron, hulk, whatever).

(Note: with things like iterons and other smaller transport vessels, a margin of error would need to be added to this study to make up for those who are not necessarily moving ore from belt to station, but are just traders.)

OK, so you got your database. Now, you calculate, on average, the amount of ships destroyed in each systems (ships, again, being mining vessels) for each cohort over a set amount of time. Let's make it...oooh a month? Combine results for the four systems and you should be able to create an average based on your results that can be counted as the probability of a mining vessel being ganked in highsec, low sec, and nulsec.

If you're really bored, you can break statistics down per ship type. At least I THINK that's how you could get the probability you were talking about.