Recommend
3 
 Thumb up
 Hide
16 Posts

BoardGameGeek» Forums » Gaming Related » General Gaming

Subject: Multiple Title Euro Game League Scoring rss

Your Tags: Add tags
Popular Tags: [View All]
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
Greetings,

I'm organizing a league to begin in Metro Detroit in 2017 of the 5 "best" heavy euro games of 2015, and I'm trying to figure out how to score each of the games played in the league.

The struggle comes when trying to determine how to give out points for second and third place finishes. In a casual gaming situation there's never any glory in second place, so we have to decide whether or not to score the game in such a way that players are rewarded for finishing in second. Obviously we're also compelled to look at players scores at the end of the game to see if any information can be gleaned from them as compared to scores from other games(regardless of where they placed in the current game).

I have so far considered the following methods for awarding league points based on game finishes.

1) Winners only. Whoever has the most wins at the end of the season wins, if there's a tie then there's a playoff game. PRO: This is attractive because most games are designed to determine only a winner and maybe not necessarily second and third. This would promote play as everyone is used to outside of the league format - simply try to win the game. CON: This will not capture above average performances of the losing players. Also, there is some "match-up" hazard where maybe the 3rd best Gallerist player happens to play one game with the 1st and one game with the 2nd best player, and the 7th best player gets a win because they happened to play 8th, 9th, and 10th.

2) First/Second/Third/Fourth get 4/3/2/1 pts - or similar. PRO: This is the seemingly simplest way to work the scoring in such a way that good performances are acknowledged without necessarily winning. The 3rd best Gallerist player from the scenario above might score 2 seconds for 6 points. If the 7th best player wins against 8, 9, and 10, then they'll likely come in 4th in their other game for a total of 5 points. CON: I think this method comes with the most "game integrity" hazard. Let's say on the last turn of a game I have 2 options. Option 1 is risky - Draw a face down card with a 25% chance of drawing something that fits my engine and would give me 20VP and most likely win the game, but 75% chance I'll get 0vp and probably end in 3rd or maybe even 4th. My second choice is to take a face up card worth 5vp, but would be taking away 15vp from another player who would likely come in second. Even though the 5vp gives me a 0% chance to win outright, I move to a near 100% chance of taking second. I think in this scenario outside of the league format no one would pass on the chance to win the game in order to screw the player in second, but with 4/3/2/1 scoring, screwing second would absolutely be the correct play. Some of this would be mitigated by something like instead of 4/3/2/1 award scores like 6/3/2/1 or 6/3/1/0.

3) Points based. Award players points based on their score compared to all other scores. PRO: Similar pros to the method above - gives due credit to above average losers. This would also maintain or maybe increase the integrity of the game for the poorly performing players. If I tried a strategy early that failed maybe I know I'm going to lose, but with this method there's still incentive to score as many points as possible, and unlike method #2 there's no incentive to sabotage 2nd or 3rd place. Outside of a league environment if I'm losing I might tune out or play in such a way I think it will make the game end more quickly. CON: Employing this method alone would probably undervalue a win. Also, there might be some integrity hazard of "running up scores" where players opt to not end the game when they're winning in order to try to score 30 more points on their next turn.

In selecting a methodology for our league I considered the following tenet most applicable. Minimize the effects of the league scoring structure on in-game decisions. i.e. Don't change the way the game is played.

For the sake of determining a method I assumed all 4-player games. I've proposed a 50/50 blend of 1 & 3 from above. I think that a "point step" (by which I mean a relatively large increase in league points based on a potentially small increase in game points) anywhere except for to reward 1st place would cause too many game integrity issues (as outlined in #2). The goal with adding score-based points would be to account for strong showings by players which didn't result in a win. My proposal was something like this:
1) Winner gets 2pts
2) All scores from all games played (grouped by title) are assumed to be normally distributed, and the results are scaled from -0.5 to 1.5 points for each player (this means players on average receive 0.5 points, and since these are 4-player games an average of 2 points are awarded per game, same as for winning).

After airing out the proposal a couple issues arose.
1) Does the scaling you chose put too much focus on points and not enough on winning?
2) What about for games where scores can't be compared from one game to the next? High scoring games for some titles are more dependent on random factors than on player skill.

So, I've so far created 3 models to look at these issues.

Model 1 (Original Proposal): 2pts for winner, -.5 to 1.5 based on total points.
Model 2: 2pts for winner, 0.0 to 1.0 based on total points.
Model 3: 2pts for winner, 0.0 to 1.0 based on total points, normalized across games (so the extra 2 points will basically land in each game as opposed to across all games)

I applied the results of these models to a fake data set and attached the results. There are 10 hypothetical players that played this title twice each for a total of 20 plays. I added two hypothetical player names so I could get a sense of how they were being scored. Guy has achieve lower scores in the games he played, but he won one of his games with a score below the average score across all 5 games. Dude on the other hand has put up 2 very high scores, however, he has been beaten twice none-the-less.

I created this thread at first for league players to comment but I'm open to anyone's input. Thanks!

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Matthew Eder
United States
Canton
Michigan
flag msg tools
mbmbmbmbmb
kysweene wrote:
I think this method comes with the most "game integrity" hazard.


While I think this is true, I have a hard time considering doing something in-game that gives me a lower chance to win the game just to possibly earn a couple of league points.

Not sure about some of the other jerks we play with though.
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
mulliganm wrote:
kysweene wrote:
I think this method comes with the most "game integrity" hazard.


While I think this is true, I have a hard time considering doing something in-game that gives me a lower chance to win the game just to possibly earn a couple of league points.

Not sure about some of the other jerks we play with though.


I wouldn't rule it out. But even still, do you think a 4/3/2/1 type model would be better than something points-based?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmbmb
If you're assuming all 4-player games, then I'd simply give 1 point to the winner and 0 points to everyone else. In case of a tie for first place, then the n tying players get 1/n points each.


I'm not a fan of point-based awards because in many games, your points don't necessarily reflect how well you played (e.g. perhaps you did a clever risky gambit that had a good probability of making you win but some chance of catastrophe - and you got unlucky). And it could encourage people to play for secure second place close to 1st place instead of trying to win.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls

Michigan
msg tools
badge
Avatar
mbmbmbmbmb
I am very intrigued by this discussion.
I have been wanting to do something like this for my game group.
My initial thought was the 4/3/2/1 method and then have a playoff game with the top four players as a final winner takes all game.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Anon Y. Mous
msg tools
mbmb
Either give 1 point to the winner and 0 to everyone else, or base points on position (3 to the winner, 2 for second place, 1 for third, 0 for fourth). Ignore in-game scoring entirely.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
russ wrote:
If you're assuming all 4-player games, then I'd simply give 1 point to the winner and 0 points to everyone else. In case of a tie for first place, then the n tying players get 1/n points each.


I'm not a fan of point-based awards because in many games, your points don't necessarily reflect how well you played (e.g. perhaps you did a clever risky gambit that had a good probability of making you win but some chance of catastrophe - and you got unlucky). And it could encourage people to play for secure second place close to 1st place instead of trying to win.


I agree with only awarding discrete points to the winner, but none of the methods I proposed at the bottom of the post encourage people to play for second, and since this is a league, risky-gambit style play should be scored appropriately over the course of 10 games. I'd agree that sometimes total points don't represent how well a player played, but "better players score more points" must be true most of the time.

Also, I should say especially to those that intend to play - I use comparative language like "better players" and "player's skill" rhetorically quite often in these posts, but this shouldn't be taken literally. I don't imagine that we're going to capture the skill level of every player and it is of course possible for good players to have bad scores and bad games.

I'm trying to create a model that approximates who the best player is based on a limited data set, and my instinct is that players that routinely score more points (and win games) are probably good players.

1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brad Keusch
United States
Ann Arbor
MI
flag msg tools
badge
It's hard to remember we're alive for the first time. It's hard to remember we're alive for the last time.
Avatar
mbmbmbmbmb
I'd like it to start at like 1500 and fluctuate from there based on performance like MMR ratings
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ryan Zombo
United States
Macomb
Michigan
flag msg tools
badge
Avatar
mbmbmbmbmb
Definitely interested in the league no matter the scoring mechanism chosen, so thanks Kyle for the idea.

I am a fan of keeping things simple and I think most of the people in our group are pretty cool and won't abuse a simple system, so I'd go with something people can understand easily like 4/3/2/1 or something where I know how many points I'm going to score in a game and I don't need a calculator or doctorate in stats to figure out. You also don't want to make it too hard on yourself to figure out scores.

That is my $0.02.

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ryan Malmberg
United States
Michigan
flag msg tools
Avatar
mb
I would use NemeStats for ease. May not be perfect for a tournament, as it rewards participation while an Elo-based system would not.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
anatana wrote:
I'd like it to start at like 1500 and fluctuate from there based on performance like MMR ratings


That's how the ratings will work, which will affect handicaps for future leagues, but starting with all new players I'll calculate a seed score at the end off of the first 10 games as though everyone was new each game.

ELO/MMR ratings are more "mathy" but pretty straight forward. I'll share how I do them but I won't gather input.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
malmby wrote:
I would use NemeStats for ease. May not be perfect for a tournament, as it rewards participation while an Elo-based system would not.


I'd prefer to retain control over all scoring algorithms and formulae.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
zomborya wrote:
Definitely interested in the league no matter the scoring mechanism chosen, so thanks Kyle for the idea.

I am a fan of keeping things simple and I think most of the people in our group are pretty cool and won't abuse a simple system, so I'd go with something people can understand easily like 4/3/2/1 or something where I know how many points I'm going to score in a game and I don't need a calculator or doctorate in stats to figure out. You also don't want to make it too hard on yourself to figure out scores.

That is my $0.02.



I agree with you and Matt that in a hypothetical scenario where everyone at the table agrees to ignore the league score and play a straight forward game intent on winning then there will be no (integrity) problems. And I think with 4/3/2/1 any "match-up" hazard would work itself out over 10 fairly scheduled games.

I guess my challenge is to imagine playing with a min/maxing type gamer (not that hard to imagine). In fact, I think having only played casual games with our group I'm not sure that in a league environment with 4/3/2/1 scoring some of the group wouldn't instantaneously change to max their league score at the sake of the game's integrity. And, as soon as one player openly began making plays like this, I'm sure everyone would follow. I would certainly be playing for league score from the get-go.

One of the biggest hazards I see comes from someone gaining perhaps a really solid foothold and commanding a strong position early on. If I'm in third or fourth place, why waste a single move trying to stop the player in first? Personally I would be kinda bummed where the game is basically conceded to 1st early and if I'm in second place, I have to be more worried about 3rd and 4th place trying to prevent me from scoring that I do from 1st place.

So, as long as going into it everyone is clear that games will be played differently I'm fine with it. Part of me is just positive someone's going to complain "well I was in a good position to score a lot of points on the last turn but then Tom took the last stone even though he didn't need it, so Bill finished in first by a mile and I ended up 3rd." and blame the fact that it's a league vs. that the scoring encourages this.

I'm not worried about the legwork of a different method. I'd just send out an excel sheet with updated scores every week. Once the formulae are decided on I'll literally just be entering scores.


EDIT: Also wanted to note that I won't be playing in case a game ever needs a sub. Also, I decided on all 4 player games because most of these titles max at 4 and are best with 4 and it makes sceduling easier. At first I wanted to include more 3 & 5 player games, but the only title that supports that is Food Chain Magnate, but then I wondered if the 5 player game might just take too long.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Matthew Eder
United States
Canton
Michigan
flag msg tools
mbmbmbmbmb
I forgot FCM was one of the games.

Not sure if you've played, but part of the game is that the players don't know how long the game goes. So if everyone adds $100 to the pot, the game is fairly short but if everyone adds $300, the scores will be MUCH higher.

Just an FYI.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kyle Sweeney
United States
Michigan
flag msg tools
Avatar
mbmbmbmbmb
mulliganm wrote:
I forgot FCM was one of the games.

Not sure if you've played, but part of the game is that the players don't know how long the game goes. So if everyone adds $100 to the pot, the game is fairly short but if everyone adds $300, the scores will be MUCH higher.

Just an FYI.


OK, I'm going to learn all of the games by the time the league starts so
I'll be able to adjust. The third column of the scoring I put in the OP would adjust for large discrepancies between games (no matter how large).

Does anyone want to play something Sunday morning at my place?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brian Hacker
United States
Michigan
flag msg tools
mbmbmbmb
I'm not terribly worried about the methodology, so long as we can trust Kyle to keep track of things and keep it honest. Frankly, the appeal to me is not my standings at the end of the league, but rather the experience of playing a bunch of new games with a bunch of fun people.

That being said, min/maxing is a thing, and some people will probably really enjoy the fine details of the point standings. If that's the case, more power to you. I'm not going to quibble about the rankings, or points-per-play, or what have you.

so my vote is whatever everyone else wants to do.

2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.