No, you do not understand why your approach is wrong.
Yes, but a random person from a pool of potential killers is also 6 times more likely to be white. So it's a wash.
No, it isn't a wash. That has zero impact on these stats. The stats are unaffected by the probable race of any killer. They reflect the relative % of potential white and black victims that any killer of either race is likely to encounter, which is the same ratio no matter how many killers there are. The question is not who is more likely to be killed by someone of the other race, b/c that is heavily a function of simple population sizes and doesn't reflect systematic targeting of the other race. The question is, when someone of each race kills someone, are they differentially targeting the other race beyond other factors that should equally impact other race killings in both directions (random probability and proximity).
Black people constantly travel through, shop in, work in areas with a large % of whites who are the majority in every random direction from their home. Whites rarely travel trough, shop in or work in areas that are a large % blacks.
I don't agree that white people rarely "travel trough, shop in or work in areas that are a large % blacks" because most, if not all, major cities have "a large % [of] blacks".
And many white live in areas where finding a single black person for 20 miles is difficult.
20 mile radius is ~1250 sq. miles. Even in Vermont, you are liable to find some black people in such a big area.
And relatively few white people live in areas such as Vermont. Most white people live in cities and suburban areas that are at least somewhat diverse. Even Burlington, VT is 5% black.
No black person in the US lives in an area where it would take more than 5 minutes to find a white person or more than 15 to be surrounded mostly by whites. Thus, just by random probability, the % of victims killed by blacks who are white should be 6 times greater than the percentage of people killed by whites who are black. But it isn't, it's only double which is 1/3 of what is should be. Thus, victims of whites are 3 times more likely to be black than they would be if they were only targeting blacks at the same rate as blacks were targeting whites.
Again, even if all that were true (and it is not) you keep hitting the same problem that it applies both to the pool of potential victims and of the perpetrators. Therefore, it cancels out and what remains, and cannot be conjured away, is the fact that twice as many black people kill white people than vice versa.
No, it doesn't cancel. Imagine you have a black and a white killer, each moving about and you have 100 black people and 600 white people also moving about. Ignoring segregation-caused proximity, what are the random odds that each killer kills a white man? It is 6 out of 7 or 6:1 in ratio terms. Both should kill 6 times more whites than blacks if their killings are random.
If we double the # of killers to 2 of each race, the answer is the same. If we say there are 6 white killers and 1 black killer, the answer is the same.
But the observed ratios are actually 11:1 to for white killers and 1:5 for black killers. So, both group of killers are killing fewer people of the other race than predicted by random chance. And this makes sense b/c proximity is a major determinant of who a killer kills. But proximity works equally both ways, and this is where you are likely getting confused. Segregation means that more that blacks are clustered and live around mostly other blacks, the more that whites live around mostly whites and the fewer blacks they encounter. However, the data show that proximity is not an equal determinant of who is killed by each race. For black killers, this proximity is leading them to killing 30 times fewer whites than predicted by population demographics, 6/1 ÷ X = 1/5, solve for X = 30. Therefore, a similar size effect of proximity for white killers mean 6/1 * 30 = 180/1 (since the proximity works changes the chance ratios in opposite direction for each race of killer, the ratio is divided by X for blacks and multiplied for whites)
In other words, if white and black killers where equally targeting victims via a combination of random population probability and proximity, then the observed 1:5 white:black victim ratio for black killers would be mirrored by a 180:1 white:black ratio for white killers. Instead, we see a 1:5 ratio for black killers but a mere 11:1 ratio for white killers. That means that whatever factors beyond population demographics and proximity that are leading to other-race killings are 15 times stronger among white killers than black killers. And your meaningless stat you are referring to is nothing but you engaging in the base-rate fallacy.
