Hi All,
Emerging from the woodwork to announce a (loooooong time comin) paper (coauthored with alpha and two others) that is now on the ArXiV (think IMDB for research papers, without all that peer review nonsense)
http://arxiv.org/abs/1512.04333
In it, we prove that, in ordinary risk, the attacker needs approximately 85.3% the number of defending armies to have a 50% chance of "essentially conquering" an enemy territory. This is probably *approximately* well-known to most veterans of this site. Our contribution is in proving that conquer odds transition from low chance to high chance of conquering over a small number of armies relative to the total armies in combat. Our paper makes all of this precise.
in general, the 50/50 cut-off occurs when the attacker has (2-avg)/avg*D where D is the number of defending armies. For ordinary RISK, the attacker expects to take about 1.08 defending armies in a single attack an average. Thus
A = (2-1.08)/1.08D = about .853D for the 50/50 cutoff.
Past papers use Markov Chains to get exact conquer odds. We omit the low unit battles to get the cut-off, which should absolutely inform every large-unit attack you make in ordinary RISK.
Much of this is by no means new, and by no means ruins RISK but simply an homage to many blissful hours wasted on a RISK clone site- and that these sorts of papers actually have a chance of getting published. It's also shameless self-promotion of my new site.
coreymanack.com
which mostly explains my retirement.
Hope all is well team.
-Mongrel
Damn. That's pretty legit.
Can you cite you own work if it's still in review? j/k
Nice work!
Very cool writing.
Sidebar - I'd never heard of Franklin and Marshall College before last week, when I was in Lancaster for the first time for work. We had dinner at Iron Hill Brewery and we walked around the campus after that. Beautiful little city you have there. Didn't know it was home to WarGear royalty.
Mongrel wrote:Hi All,
Emerging from the woodwork to announce a (loooooong time comin) paper (coauthored with alpha and two others) that is now on the ArXiV (think IMDB for research papers, without all that peer review nonsense)
http://arxiv.org/abs/1512.04333
In it, we prove that, in ordinary risk, the attacker needs approximately 85.3% the number of defending armies to have a 50% chance of "essentially conquering" an enemy territory. This is probably *approximately* well-known to most veterans of this site. Our contribution is in proving that conquer odds transition from low chance to high chance of conquering over a small number of armies relative to the total armies in combat. Our paper makes all of this precise.
in general, the 50/50 cut-off occurs when the attacker has (2-avg)/avg*D where D is the number of defending armies. For ordinary RISK, the attacker expects to take about 1.08 defending armies in a single attack an average. Thus
A = (2-1.08)/1.08D = about .853D for the 50/50 cutoff.
Past papers use Markov Chains to get exact conquer odds. We omit the low unit battles to get the cut-off, which should absolutely inform every large-unit attack you make in ordinary RISK.
Much of this is by no means new, and by no means ruins RISK but simply an homage to many blissful hours wasted on a RISK clone site- and that these sorts of papers actually have a chance of getting published. It's also shameless self-promotion of my new site.
coreymanack.com
which mostly explains my retirement.
Hope all is well team.
-Mongrel
For the record... time here was not wasted... lets say fruitfully misspent.
BorisTheFrugal wrote:Very cool writing.
Sidebar - I'd never heard of Franklin and Marshall College before last week, when I was in Lancaster for the first time for work. We had dinner at Iron Hill Brewery and we walked around the campus after that. Beautiful little city you have there. Didn't know it was home to WarGear royalty.
Take home growlers at iron hill.