Bad Matchmaking?

Report here the bugs you experience in Blood Bowl 2.
Rydis
Posts: 8
Joined: 21 April 2015, 01:05

Bad Matchmaking?

Postby Rydis » 30 April 2017, 04:30

Been trying to play on official ladder and been experiencing really awful matchmaking. Every new team im making is just going against 1400+ TV teams. No game has been even close to fair. Last team, fresh TV 1000..rolled into a 1400 TV team. I dropped to 800 TV..next match, 1500.

Its getting a little ridiculous at this point how awful the matchmaking is. Im 11 games into this and every match has been a minimum of 400 TV difference.

I know some will come with the argument, well would you rather it find no game?

Yes..I would rather it tell me no matches to be found them keep facing such unbalanced games.

User avatar
VoodooMike
Posts: 1614
Joined: 14 July 2009, 07:44
Location: Calgary, Alberta
Contact:

Re: Bad Matchmaking?

Postby VoodooMike » 30 April 2017, 06:50

Its getting a little ridiculous at this point how awful the matchmaking is. Im 11 games into this and every match has been a minimum of 400 TV difference.
You're getting what you're getting because that's who's playing. It's not giving you bad matches when closer matches are available, there just isn't anyone around that is closer. If you're playing in CCL rather than COL then you should bear in mind that it's the competitive ladder, and the people serious about competing start their teams early... if you start a fresh team late in a season you should absolutely expect to face higher TV teams constantly.
Yes..I would rather it tell me no matches to be found them keep facing such unbalanced games.
I'm curious as to what you think "such unbalanced games" mean? Lets look at the first three seasons of CCL... games played by teams under 1100 TV against teams at least 400 TV above them (that are played to completion):

CCL Season 1 - 40.1% won (382 Games)
CCL Season 2 - 38.7% won (558 Games)
CCL Season 3 - 35.2% won (445 Games)

So what exactly is "balanced" to you? A straight up 50% win rate? Even in the worst of those seasons the difference between that and the actual win rate of these "such unbalanced games" is only 14.8%... in COL, the non-competitive ladder, the win%s for teams are higher, with the win% for teams 500 TV down rarely dips below 40%.

You're losing these games in your head, not on the pitch.
Friendly Reminder: Correlation does not equal Causation - tattoo it on the inside of your eyelids if it'll help.

JimmyFantastic
Posts: 495
Joined: 28 February 2012, 21:12

Re: Bad Matchmaking?

Postby JimmyFantastic » 02 May 2017, 14:47

Lets look at the first three seasons of CCL... games played by teams under 1100 TV against teams at least 400 TV above them (that are played to completion):

CCL Season 1 - 40.1% won (382 Games)
CCL Season 2 - 38.7% won (558 Games)
CCL Season 3 - 35.2% won (445 Games)
Now the same stats including concessions pls.
Image

User avatar
VoodooMike
Posts: 1614
Joined: 14 July 2009, 07:44
Location: Calgary, Alberta
Contact:

Re: Bad Matchmaking?

Postby VoodooMike » 02 May 2017, 15:23

Now the same stats including concessions pls.
Concession games have nothing to do with actual match balance - at best they are about incorrectly perceived balance. I'm a bit wary of diving into those stats in a thread discussing actual balance.
Friendly Reminder: Correlation does not equal Causation - tattoo it on the inside of your eyelids if it'll help.

Azathot
Posts: 1
Joined: 02 May 2017, 15:31

Re: Bad Matchmaking?

Postby Azathot » 02 May 2017, 15:53

CCL Season 1 - 40.1% won (382 Games)
CCL Season 2 - 38.7% won (558 Games)
CCL Season 3 - 35.2% won (445 Games)

So what exactly is "balanced" to you? A straight up 50% win rate? Even in the worst of those seasons the difference between that and the actual win rate of these "such unbalanced games" is only 14.8%... in COL, the non-competitive ladder, the win%s for teams are higher, with the win% for teams 500 TV down rarely dips below 40%.

You're losing these games in your head, not on the pitch.
This argument seems somewhat weird to me, since it comes from the same people that implemented TV+ matchmaking. Giving coaches with worse records a modifier in TV for their matchmaking only makes sense, if you think that having a TV advantage also improves your chances to win. Or what else would be the point of TV+ matchmaking?
Now the same stats including concessions pls.
Concession games have nothing to do with actual match balance - at best they are about incorrectly perceived balance. I'm a bit wary of diving into those stats in a thread discussing actual balance.
That is an extremely misleading argument as well.
Yes, concessions most likely include a large set of people that just can't be bothered to play out the match they anticipate to lose. But they likely also include a large set of games where you get banged on in a few turns to a point that the coach down TV barely has a chance at doing anything for the remainder of the match and opts out. If you want to correct the data by anything, then mask the games that have been played for less then 2 turns or something in that general area.
But even pre-match concessions have a strong bias, because players are less likely to concede if they see a matchup they see good chances. Just as examples (overexageration to make my point clear), take a TV1000 Orcs team vs a 1400 Nurgle team with 3 Clawpombers and Dirty Players as opposed to a 1200 Wood Elf team vs a 1600 High Elf team that is bloated by having 8 rerolls and Strong Arm first skill on every player - I would argue the Orcs are more likely to concede instantly or early, whereas the Wood Elfes probably have the edge in that matchup and by excluding concessions without any further modifiers you are skewing the data towards the more favourable matchups, which obviously results in a higher than expected win rate for the team down TV.

PS: Not related to the actual match balance, but since Blood Bowl is supposed to be a game: Another point worth to consider is that a large portion of the player base just simply does not enjoy when they are matched with such large TV discrepancies and I can personally sympathize with that because those kind of games are rarely fun, even if you win them.

User avatar
VoodooMike
Posts: 1614
Joined: 14 July 2009, 07:44
Location: Calgary, Alberta
Contact:

Re: Bad Matchmaking?

Postby VoodooMike » 25 May 2017, 01:38

This argument seems somewhat weird to me, since it comes from the same people that implemented TV+ matchmaking. Giving coaches with worse records a modifier in TV for their matchmaking only makes sense, if you think that having a TV advantage also improves your chances to win. Or what else would be the point of TV+ matchmaking?
Higher TV does give you an advantage, but the point of showing you the win rates was to demonstrate that, even at the highest TV differences, the amount of advantage does not appear to be overwhelming. The win rate difference is around 10% in COL... so anyone who claims the "large" TV differences are unbeatable is simply wrong... they're losing the psychological battle.
That is an extremely misleading argument as well.
It's not misleading... declaring it to be is miselading unless you have contradictory data to present.
Yes, concessions most likely include a large set of people that just can't be bothered to play out the match they anticipate to lose. But they likely also include a large set of games where you get banged on in a few turns to a point that the coach down TV barely has a chance at doing anything for the remainder of the match and opts out. If you want to correct the data by anything, then mask the games that have been played for less then 2 turns or something in that general area.
Not necessary to "correct" the data. We don't know why people concede, and assuming that any match played for a certain amount of time means a concession is for a good reason rather than a dumb reason is disingenuous.

In CCL you are only permitted 5 concessions before you are tossed out of the league for the season. People playing most of a match and then conceding rather than losing, when concession not only gives you a loss but also potentially gets you banned from the league... seems to preclude "a large set" of those being justified choices... especially since most of the concessions in CCL belong to people who subsequently DO get kicked out of the league.
But even pre-match concessions have a strong bias, because players are less likely to concede if they see a matchup they see good chances.
Which brings us right back to my point: you're losing those matches in your HEAD... you've decided you'll lose and then you make it a reality. The stats don't support the idea that these are unwinnable matches, but you lose every match you refuse to play.
PS: Not related to the actual match balance, but since Blood Bowl is supposed to be a game: Another point worth to consider is that a large portion of the player base just simply does not enjoy when they are matched with such large TV discrepancies and I can personally sympathize with that because those kind of games are rarely fun, even if you win them.
A "large portion", eh? What percentage is that? I mean.. if you're saying it you must have some data on the topic, right? No? Didn't think so. I think we should stick to concrete facts and figures instead of inventing new ones.
Friendly Reminder: Correlation does not equal Causation - tattoo it on the inside of your eyelids if it'll help.

JimmyFantastic
Posts: 495
Joined: 28 February 2012, 21:12

Re: Bad Matchmaking?

Postby JimmyFantastic » 25 May 2017, 10:35


CCL Season 1 - 40.1% won (382 Games)
CCL Season 2 - 38.7% won (558 Games)
CCL Season 3 - 35.2% won (445 Games)
Is this win% ? or Just % of wins? What does it look like for even TV?
Image

User avatar
VoodooMike
Posts: 1614
Joined: 14 July 2009, 07:44
Location: Calgary, Alberta
Contact:

Re: Bad Matchmaking?

Postby VoodooMike » 25 May 2017, 11:22

Is this win% ? or Just % of wins?
Those are the same thing.

If you're trying to ask what the long-term win%s for each individual coach are, then combined... that's nonsensical. The numbers presented there are for matches in which the maximum TV difference existed (500). There's no coach who plays ALL their games at maximum difference... unless they only play one game in which case their "long-term" win% is 0, 50, or 100... and not useful to the topic.
What does it look like for even TV?
Identical TV? That'd be 50%.. again, by definition.. since both teams would be included in that dataset, for every game, and in each game one would win, one would lose, or both would draw...
Friendly Reminder: Correlation does not equal Causation - tattoo it on the inside of your eyelids if it'll help.

JimmyFantastic
Posts: 495
Joined: 28 February 2012, 21:12

Re: Bad Matchmaking?

Postby JimmyFantastic » 25 May 2017, 11:28

Win% in Bloodbowl counts draws as half a win and half a loss. So this % of games won you had was actually win% and not % of games won. It's an important distinction because in the NAF Wood Elves have a 56 win% but only actually win 45% of the games.
40-35% win% is clearly awful, Halflings have a 36 win% in the NAF...
Image

User avatar
VoodooMike
Posts: 1614
Joined: 14 July 2009, 07:44
Location: Calgary, Alberta
Contact:

Re: Bad Matchmaking?

Postby VoodooMike » 25 May 2017, 12:16

Win% in Bloodbowl counts draws as half a win and half a loss. So this % of games won you had was actually win% and not % of games won. It's an important distinction because in the NAF Wood Elves have a 56 win% but only actually win 45% of the games.
40-35% win% is clearly awful, Halflings have a 36 win% in the NAF...
I don't think it's an important distinction in the long run, personally.. especially in this case since we're discussing whether large TV differences result in "unbalanced" games. If you manage a draw then either the game was sufficiently balanced in terms of mechanical ability, or coaching skill.

Even if you think that what you're calling "% of wins" is important, it requires context in that it needs to be compared to that roster's "% of losses", as well as those figures of other rosters. This is a game where you CAN have draws, and deciding to ignore those when examining balance is... strange.
Friendly Reminder: Correlation does not equal Causation - tattoo it on the inside of your eyelids if it'll help.


Return to “Technical forum”

Who is online

Users browsing this forum: No registered users and 1 guest

cron