Oddly enough both point systems converted to odds give me around 2.5 out of 10 wins against Jeep, which is 99.9999% not happening.

Your 10 games against him are biased because they are only Strat 1v1. Looking over your CW profile, I see you mixed in a lot of SEAD. When did Beep Beep Jeep last play CW... Ah, just opened up a random season with 2 games from him. It was SEAD and Bomb ;)

Your 10 games against him are biased because they are only Strat 1v1. Looking over your CW profile, I see you mixed in a lot of SEAD. When did Beep Beep Jeep last play CW... Ah, just opened up a random season with 2 games from him. It was SEAD and Bomb ;)

It still shows a clear flaw in my system. It wrongly assumes that I have 25% chance against it, and if we'd both que for S1V1, it would still think I have 25% chance against him. Of course I will just ignore the experiment anyway because a sample size of 10 is meaningless, but I wouldn't claim my ratings are perfect, or that 5s ratings are perfect either. Good enough though - Mediocrates.

Edited 9/29/2023 07:20:45

Hm, I believe you are missing what the numbers are showing. The numbers do not show that you have a 25% chance "in Strat" against him. How about you specify in the bomb? With some practice I believe you might be able to overtake him in his rating, but you still won't be able to get your 25% strat win rate.

Edited 9/29/2023 07:27:26

Hm, I believe you are missing what the numbers are showing. The numbers do not show that you have a 25% chance "in Strat" against him. How about you specify in the bomb? With some practice I believe you might be able to overtake him in his rating, but you still won't be able to get your 25% strat win rate.

No no no... That's not what I was saying. I was saying that a matchmaker based on player ELO would assume that win rate on any map because those are the ELO points of the players. Therefore it would create some mismatches. The solution would be to calculate a per map ELO, but the issue there that it would be even worse because of low number of games.

edit: So there is an actual question here: is a per map ELO based on fewer games more precise in predictions than a per player ELO. This is something that just got added to my test todo list.

Edited 9/29/2023 07:32:31

I personally prefer 5s ratings to mine because I am #289 on his list and all the way down at #504 on mine. Oddly enough both point systems converted to odds give me around 2.5 out of 10 wins against Jeep, which is 99.9999% not happening.

The meaning of this is probably that you have improved a lot recently. I mean you are now number 42 on the 1v1 ladder which does not fit with your CW rating being bottom feeder tier.

Edited 9/29/2023 08:58:26

The meaning of this is probably that you have improved a lot recently. I mean you are now number 42 on the 1v1 ladder which does not fit with your CW rating being bottom feeder tier.

Nah, I'm hopeless, but I was joking in that post either way. The main difference between the 2 systems is that mine rates alts much higher than his does, and since there will be plenty of temporary alts, most people will end up lower on my ratings compared to 5s.

edit: When you're trying to find fair matches for people it's way better to have a person or alt with 15/0 score up at position #19 than at position #265. But if you're trying to make some sort of leaderboards, or reward people based on those rankings, then it's better to keep him at #265 instead. That is why I think my system is better for matchmaking and 5s is better for just making leaderboards. You shouldn't use a system for matchmaking that gives a 15/0 W/L player 278.9 points and a 64/51 W/L player 269.2 points. There is clearly a much wider gap between those 2 players.

Edited 9/29/2023 09:12:36

The series concluded 9-1 with a stupid loss from me in the last game.
GG's

https://www.warzone.com/MultiPlayer?GameID=35738873
https://www.warzone.com/MultiPlayer?GameID=35738977
https://www.warzone.com/MultiPlayer?GameID=35739012
https://www.warzone.com/MultiPlayer?GameID=35739130
https://www.warzone.com/MultiPlayer?GameID=35739246

We conclude that beep can beat himself in 1 out of 10

We conclude that beep can beat himself in 1 out of 10

I am pretty sure I could get that 25% win rate against him if someone would smash a frying pan to the nearest wall next to him to annoy him while he's playing. Or if someone would tickle him. Either way, I had fun.

I hope you correctly included frying pans and tickling into your elo calculation

The blockade seen around the world :o

Well done spider! 1 win against beep is still pretty good :p

Question is ... what happened on T18? Did Orange play the turn of his life or did BBJ fall asleep for 1 turn? (OSvBBJ) 19v18 income at start of turn with 83v105 army counts after deployments, but then two 5-income bonuses get broken, ending with 19v8, ouch! (still an army lead tho, but with an income gap that big, Orange should still easily win that at that point)

Actually I see BBJ didn't defend either 5-bonus (I'm ignoring the +2 to Atlantic NE b/c there was already 7 on Quebec, already enough to take both GL & E.US sending 3 each way) and deployed in WR & Asia to prevent WR & WC from being taken

Either way, take the glory Orange ... 1 in 10 or 1 in 1,000,000, it's still 1 win which is more than most of us have

Edited 9/29/2023 13:57:49

I think the problem with this thread is that anyone that can make a change to the CW system gives a flying F about it.

+1

i can’t remember who said it earlier in the thread comparing it to the clan league thread. the amount of chatter here by active wz members shows we care. we don’t agree on the problem, let alone the solution but we’re spending tine trying to mcwga (make clan wars great again) was it ever, i think it still is otherwise there wouldn’t be 8+ pages of posts.

i can’t remember who said it earlier in the thread comparing it to the clan league thread. the amount of chatter here by active wz members shows we care. we don’t agree on the problem, let alone the solution but we’re spending tine trying to mcwga (make clan wars great again) was it ever, i think it still is otherwise there wouldn’t be 8+ pages of posts.

True as that might be there is only 1 person who actually can change anything and that person isn't here, now is he?

for those that weren’t there. in jun/july ‘22 there was a push to make changes to cw. ever interested clan sent a delegate and there is still a google doc with the ideas. it fizzled because we all have rl and this is a project that requires dedication. the proposals, once ironed out and agreed upon more than less would have been presented to fizzer. i can link the doc later today when i’m on pc.

^as old yeller said, we talked a lot about it last year ... and realized that not enough people agree on any single solution, so there's not even really a unified backing to request a clear change, so fizzle it did

why does this matter? b/c the best way to get something changed is to get a clear & unified request with 600+ players backing it onto uservoice

but in the end, everyone has different ideas on what's required to make CW better; even something that to me seems ridiculous such as adding Duel to the CW templates (imho if we're going to do that, I'd rather just decomm CW altogether and give out the rewards as raffles in GC instead, b/c that's essentially the same thing so let's stop wasting time play "games" (Duel isn't really a "game" imo)) but it wasn't just 5s arguing for it, three were others, and this is just 1 of many points discussed, and very few agreed on much that would be materially changed CW, so in the end, I realized if we can't even agree on what "good" is, so there's no point in discussing further, b/c we're working in totally different directions

better said than i did krinid

that person isn't here, now is he?

He just went out for some milk and a pack of smokes. Momma says he'll be back any day now