The idea behind ranked choice voting is that your 2nd choice candidate has a chance, but round 1 eliminates anyone who doesn't receive a 1st place vote. If everyone agrees that Smith is the 2nd best candidate in a huge field, he will get eliminated in round 1. That's not fair! I have solved that problem with my voting method. Round 1 should eliminate the biggest loser and I determine that by examining everyone's ranking against each other.
I can show you how to do it by hand and there is a video coming, but how would I write the software? I want to take everyone's votes and put them in a table and then analyze that table to create a new table without the favorite. Then analyze that table and eliminate the favorite to create a new table.... and so on until we are left with the biggest loser.
That person gets eliminated in round 1.
Then create a new table for round 2 and repeat the process until there's only 1 candidate remaining.
So you see spoilers have little to no effect. Guys like Trump who get a lot of last place votes will have a tough time. It's not winner take all; you can rank the winners so you get more meaningful data.
Where do I even begin?

wheelflush,

What you describe sounds a lot like the Australian electoral system. I imagine that there's some sort of software that is designed for that, although it is a substantially manual process. Hopefully this is helpful to you in finding a pre-existing solution/algorithm/etc.

i need to dive deep here. australia and maine use rcv, but is it instant runoff voting?
my system is more accurate than instant runoff voting. it's hard to illustrate exactly how beneficial this would be. if i knew how to code it would be so easy!

I suggest the first step should be to write and publish your plan in coherent English. A blog is probably the easiest place to start.

That I can do!

That all depends on your definition of "accurate", doesn't it?

yes and we will delve deeper using better terms like condorcet winner, condorcet loser, pareto criteria, and monotonicity.
but for now, lets just say accurate means it takes into account who you like AND who you hate.

wheelflush,

When you make your first draft of this thing, get chatgpt or claude.ai to rewrite your first draft in something that is simpler and more accessible to normal people.

I'm a hyper-interested person in political voting methods, and even I think that you sound like a crazy conspiracy theorist.

OK
i'm curious what conspiracy you think i'm pushing. i just think i found a better alternative.

No, I don't mean to say that this is a conspiracy. I'm saying that mainstream audiences will be alienated by the way you've communicated above.

For example, compare this:
to this

Obviously, edit to taste, etc etc

I see the robots are replacing me lol. Love it.
Anyway, here's the blog post I'm working on... It's probably too technical already but I got carried away.

Example 1. Consider a 3 way election between A, B & C with 100 ballots. Voters are asked to rank their choices from 1 to 3. Here are the results.

ABC = 19
ACB = 0
BAC = 36
BCA = 10
CBA = 0
CAB = 35

Who's the winner?

First Past the Post (American system)
Count your first choice only.
A has 19 first order votes. B has 46 first order votes. C has 35 first order votes.
Result. B wins. C is the runner up.

The Instant Runoff Method (Australian system)
Round 1. Eliminate the smallest favorite. A has 19 first order votes. B has 46 first order votes. C has 35 first order votes. A gets eliminated. Everyone else moves up a spot.
Round 2. B beats C head to head (65).
Result. B wins. (But who do you consider the runner up?)

The Lenhok'sin Method (my system).
This is what I want to automate. It's far too cumbersome to do it by hand. Open to suggestions
Round 1. Eliminate the biggest loser. B has a plurality of first order votes (46). Ignore B.
A beats C head to head (55). C is eliminated in last place. Everyone else moves up a spot.
Round 2. A beats B head to head (54).
Result. A wins. B is runner up.

The Minimax Method
Analyze everyone head to head and just count their worst performance. The highest vote getter is the winner.
AB = ABC + ACB + CAB = 54
AC = ACB + ABC + BAC = 55
BA = BAC + BCA + CBA = 46
BC = BCA + BAC + ABC = 65
CB = CBA + CAB + ACB = 35
CA = CAB + CBA + BCA = 45
Result. A wins (54). B is runner up.

The Copeland Method
The candidate with the most head to head wins is the winner.
A beats everyone head to head. C loses to everyone head to head. C is eliminated in last place.
Result. A wins. B is runner up.

The Borda count.
Your first choice gets 2 points, your second choice gets 1 point, and so on.... Highest score wins.
A = 2(19+2) + 36 + 35 = 109
B = 2(36 + 10) + 19 + 0 = 111
C = 2(35 + 0) + 0 + 10 = 80
Result. B wins (111). A is runner up.

The Dowdall count. (Nauru system)
Your first choice gets 1 point, your second choice gets 1/2, point, your third choice gets 1/3 point, and so on.... Highest score wins.
A = 19 + 2 + (36 + 35)/2 + 10/3 = 57.8
B = 36 + 10 + 19/2 + 35/3 = 67.2
C = 35 + 10/2 + (19 + 36) /3 = 58.3
Result. B wins (67.2). C is runner up.

Why do we care who is the runner up?


Also, if you can't explain your system to the average voter in a way that they understand in two minutes or less, no one will trust the system.

My method has applications beyond just winner take all elections. For instance say you want to rank the top N Beatles albums based on a survey of 100 people. I can do that.

That's where I need your help writing a script. Ask people to rank the top 10 films of all time and they can see the results immediately, change the variables around, share it...

Here's a quick example that makes the benefit of my method plain to see.
On Dec 3 the College football rankings came out. You can see all the votes here:
https://collegepolltracker.com/footb...d/2023/week-15
35 voters placed FSU above Texas.
27 voters placed Texas above FSU.
Yet FSU is ranked 4th and Texas is 3rd.
Couple of reasons for this. The 5th-6th place teams got between FSU and Texas. This is commonly known as spoiling. My system controls for that by eliminating spoilers in the early rounds. That way we're only comparing the relevant teams to each other.
Also, they assign points and rank the teams based on their score. It's called the the Borda system. It's completely arbitrary as you can see. My system doesn't make up scores. The votes are the scores. Just look at the votes!
The majority of the voters said Texas is worse than FSU. The results should reflect that.
I'm meeting with a programmer this week and we're going to get this started.

Instead of giving points to a combination (ABC) why not assign points based on rank: 1st place votes = 3, 2nd place votes = 2, 3rd place votes = 1 and use those to do your eliminations?

kero!!!!

very good idea! in fact i just discovered last night it has a name.
he calls it TVR.
washington post article here
which is pretty much the same theory i have been promoting. we just have a slightly different method for picking the biggest loser. still, that's pretty close!