Improvement to the auction ?

[Molfo]

@Babbuc49 <3

@Spooky If tournaments were to be introduced here on BG, the I'd like to have two parallel sessions running, one with auctions for those who believe they work and one with "last player rotate sectors" for those who don't.

[El Temblo]

@testrun said in Improvement to the auction ?:

What do you think of the idea of the last player setting values, instead of/in addition to rotating sectors? It would cut out the multiple days of "bid+1"ing you mention...

Well, it could work; but, again, it's in the hands of a single player, so... why don't we just play GP as it was intended, maybe adding just the sector rotation? :D

Quick replies: for Lantids, I think ALL of them, and no, it wouldn't make them op. Regarding Gleens, sorry, but "Gleens can already be quite strong" is just light years from my experience. Anyway, since nothing of this will take place, we're talking about nothing ;).

[El Temblo]

@Babbuc49 said in Improvement to the auction ?:

There are many ways you can tweak the auction system but the game will never be objectively better than with the good ol' rotation, just longer. It may meet your tastes better, I won't argue with that, but most of the time you'll just be playing a small gambling game before the actual game, and that's perfectly fine, I'm just not intersted in it anymore. Just mind that no auction system will improve your chances of winning a game.

THIS, a bazillion times THIS.

Out of my experience and mistakes, I now realize I could plan carefully for an auction in order to maximize my chances of winning, even with, let's say, Lantids if the mains are "planet types" and "Gaia".

I just don't want to do that (and it's fine if other people want to do auctions, ofc).
I'm definitely with Babbuc here.

[testrun]

i'm going to avoid getting into the "objectively better" rabbit hole re: "good ol' rotation". the pros/cons of rotation aren't really relevant here.

this post is about "Improvement to the auction", so it seems like the only relevant feedback is along the lines of "what could make auctions better for more players?" people may dislike auctions, or be dead set against them, and that's fine, they never have to use them. (going the full mile of rebalancing factions can easily be a parallel effort, and if last sector rotation works great then, hurray! auctions should be less appealing and effectively go away on their own...) but this thread should probably stick to constructive feedback.

the intractable issues people mention with auctioning seem to be that it's too non-deterministic / difficult to predict, and that gaia isn't supposed to be about risk-assessment.
i'd argue that every selection during setup requires a similar magnitude of prediction to bidding-- and rotating sectors a magnitude higher-- and that risk assessment is a fundamental part of every decision in the game-- am i giving power charge too soon, can i wait to take that planet, how much do i have to rush this track to get that adv tile, etc-- but: if you view those as unique or excessive issues with auctioning, they are intractable, so you should avoid playing games with any form of auction implemented.

the main potential goals of auctioning people focus on are:
G1) evening out the impact of player order re: faction selection-- OP's point
G2) evening out / making more factions playable in a given setup

the main potential flaws with auctioning people focus on are:
F1) "sabotaging" games with a bad faction*-- this was OP's key point (and most people's concern), and no one has really addressed their suggestion directly
F2) the possibility for other metagaming -- seeing what factions people are aiming for (or aiming to avoid), trying to push their prices up, etc.
F3) the delay to the game -- going back and forth on bids for days, especially with 4 players

@oelepetoetje's suggestion was "whenever someone bids zero for a tribe that he did not pick himself, he is allowed to change that tribe to one of the tribes that was not in the initial selection." @AntoineBR had a related suggestion, that factions are picked and bid on one-by-one.

• Superficially, this will address F1 in full; and partially address F2 and F3 as not being shoehorned into the "bad" faction should reduce metagaming and lower the peak value of bids, thus speeding the bidding process up.
• This will address G1, but G2 is largely unaddressed -- the relative value of a faction-with-replacement can still make the selected faction unplayable. You'll still end up with the four strongest factions in every game if people are playing (roughly) optimally
• The other concern I have (and others voiced) is that the value of a faction can change dramatically depending on what other factions are in the game, because of planet color, likelihood of gaiaforming, etc. Once people are "locked in" on another faction, they may get severely bit by a faction swap, which can be taken into account but then many of the issues with auctioning are increased-- the ability for metagaming, the difficulty of prediction, etc. You're now making more decisions with less information.

Overall, I definitely would not prefer this; but if most players only care about G1 and not G2, then something along these lines could work better overall.

My preference would still be switching to a closed bid format-- every player makes one and only one bid for all factions at the start, and then factions are assigned automatically based on the relative value of the bids.

• This completely resolves F2 and F3; there's no delay and there's no metagaming as there's no counter-bidding.
  It also completely fulfills G1 and G2: the impact of player order is now completely mitigated, and any faction can be involved
• F1, the issue of sabotaging with bad factions, is partially resolved-- those factions can still be introduced, but because bids are closed and factions assigned automatically, the person picking them has no way of guaranteeing they don't end up with the faction they tried to sabotage others with. So if you're picking it, you have to be comfortable playing it.
• The main drawback I can see for some players is that the final choice is taken out of your hands. In every other setup, you get the final decision about what faction you are, even if it's at an absurd cost. So if you just reaaally don't want to play Xenos or whoever, in the current auction you can always just keep bidding even if it's 50+ points. With closed bids, you can't absolutely guarantee that you don't get a faction (even if you can make it very likely by bidding high). How important this is will obviously vary player-to-player; I'm personally fine with it.

*as mentioned before, what people see as trying to force other players to play a bad board, I think is more often players picking factions they're interested in trying themselves; but I'm predisposed to view it that way as that's why I do it

[Spooky]

@testrun said in Improvement to the auction ?:

My preference would still be switching to a closed bid format-- every player makes one and only one bid for all factions at the start, and then factions are assigned automatically based on the relative value of the bids.

How do you resolve 'relative value of the bids'? If, for instance, one player will bid the most points for all factions than how you will determine which faction he should play? Even if you will try to create some kind of factor representing this relative value, you still may end up with tie situations. Some players may think - I can play with any of those factions, so I will bid 0 for all of them. What if there are 2 players thinking this way and in effect more than 1 player will end up with exactly same bidding values? It's not that unlikely. So you need sth extra for resolving tie situations.

Besides it's cool idea and it's addressing some issues mentioned before.

[testrun]

@Spooky good questions. There's a ton of ways to handle relative value and ties, depending on what you want to emphasize. To my knowledge, the main approaches either go for simple face value (e.g. top-down, you pay the lowest winning bid for each faction), or maximizing efficiency (true value/optimizing bids)

Example auction and bids:

Factions: Ambas/Terrans/Firaks
Player A: 10/8/0
Player B: 0/8/5
Player C: 15/10/0

The simple face value approach would basically follow a "highest overall" system, i.e.: 1. highest bid overall wins, and pays 1 more than the next highest bid for that faction; 2. remove the now excess bids for the winning player/faction; 3. repeat 1&2 for the rest. In the case of a tie, the best tiebreaker is likely to compare to the next highest valued faction, and resolve based on that.

So in the example:
Player C wins Ambas for 11 (their value was 15, but they get it for +1 over the next highest bid). Remaining bids:

Factions: Terrans/Firaks
Player A: 8/0
Player B: 8/5

Player A and B both bid 8 for Terrans, so look to their next bids.
Player A values Terrans over their next option more than Player B, e.g.: (8-0) > (8-5), so Player A would get Terrans for their max bid of 8, and Player A gets the last option Firaks for 0.

There's obviously some downsides to this, which is to say how a player values the factions in comparison to each other is actually more important than how they value them compared to other players, and that's only used as a tiebreaker above. So maximizing efficiency / optimizing bids based on game theory is probably a better way to go.

It would effectively boil down to: people pay the minimum they have to so that it isn't worth paying any more for another faction. Using the same example:

Factions: Ambas/Terrans/Firaks
Player A: 10/8/0
Player B: 0/8/5
Player C: 15/10/0

The optimal solution is likely:
Player C gets Ambas for 5; Player A gets Terrans for 3; Player B gets Firaks for 0.
(Values might change by a point depending on how you handle ties / "bidding order", so we'd need to decide a standard approach there.)

Effectively, it's as if people bet as follows:

A bid 0 on Ambas -- their highest valued faction
B bid 0 on Terrans -- their highest valued faction
C bid 1 on Ambas -- relative value 4 over next best faction (15-1=14 for Ambas vs 10-1=9 for Terrans vs 0 for Firaks)
A bid 2 on Ambas -- relative value 1 (10-2=8 for Ambas vs 8-1=7 for Terrans vs 0 for Firaks)
B stays on Terrans
C bids 3 on Ambas -- relative value 3 (15-3 vs 10-1 vs 0)
A bids 1 on Terrans -- relative value 1 (10-4 vs 8-1 vs 0)
B bids 2 on Terrans -- relative value 1 (0-4 vs 8-2 vs 5)
C stays on Ambas
A bids 4 on Ambas -- relative value 0* (10-4 vs 8-2 vs 0)
B stays on Terrans
C bids 5 on Ambas -- relative value 3 (15-5 vs 10-3 vs 0)
A bids 3 on Terrans -- relative value 1 (10-6 vs 8-3 vs 0)
B bids 0 on Firaks -- relative value 1 (0 vs 8-4 vs 5)
(Done)

* My instinct is a tie in relative value between factions for a player should be broken by their face value, e.g. if they originally valued Ambas more than Terrans, they'd take Ambas over Terrans assuming the same net value after bids. It could also be broken by which board they picked at the start, if relevant (face value could be the backup tiebreaker).

Ties between players could either be broken by player order, as above, or by defaulting to the faction they picked at the start-- e.g. if the optimal value for Terrans over other options is 3 for two players, it would go to whoever selected Terrans as an option.
(In rare cases you may end up with a true tie that can't be broken by meaningful distinctions, and then it would have to be random. e.g. if it ends up with two faction that have the exact same bids from two players who didn't pick either faction at the start, you just assign them at random since the players didn't differentiate between the two.)

It's a more complicated algorithm to implement, but doable and should get as close to optimal results as possible.

[lucasrrr]

@testrun said in Improvement to the auction ?:

F1) "sabotaging" games with a bad faction*-- this was OP's key point (and most people's concern), and no one has really addressed their suggestion directly

I did suggest this:

I think it would be nice if you could have the app pick the 4 factions that will be auctioned off randomly. That way you get to play all the permutations of factions, which is really cool (and that greater kind variety is kinda the whole of the point of having the auctions) and you could never get deliberate trash picks.

Which I still think is a solution to the "trash-picking" problem that should be pretty easy to implement (and also saves some time, as players do not have to contemplate what faction to introduce into the auction). It's how I usually play offline and would choose to play here if it was possible. It is a lot of fun to see, for example, how a game where gaia-forming is heavily rewarded develops when the factions are Lantids, Taklons, Nevlar, and all those strange combinations can come up when you have random selection of the factions + auction.

[testrun]

@lucasrrr sorry, when I said "no one has really addressed their suggestion directly", I meant their suggestion for an improvement- the option for the zero bid to swap boards

I played similarly in person too and still usually randomize when playing with friends on here (we hit random twice and pick between the two results). I definitely like seeing new combinations of factions and trying to play a faction in unusual conditions.

Re: randomizing all four boards here- this is technically possible already through coordinating in the chat, but it might be nice to formalize it. As far as whether it fixes trash picks-- the question is whether people dislike the "spirit" of it, what they see as deliberate metagaming; or whether they just don't like a bad faction being there because of how it influences the bidding / not wanting to get stuck with it. Randomizing would of course fix the first, but not the second, as there's a 4/14 chance the worst possible faction still gets selected and the same bidding war to avoid it happens anyway.

In my mind, it's possible to resolve the bidding war part of things with a closed bid system, but still preserve the ability to have unusual factions in the game.
(And these two options could always be combined-- randomizing factions would work as an independent option you could select during setup)

[lucasrrr]

@testrun ah, of course, should have picked that up from your discussion of @oelepetoetje's suggestion below that

I think a good closed bidding system would be a nice way to solve what you've termed F3) -- the delay caused by slow incremental bidding -- which I don't terribly mind, but definitely could do without (also agree that this is completely independent from and could be combined with randomizing faction selection). Do worry such a system might not be intuitive for a lot of players though.

[testrun]

I'd like to think Gaia players can handle some complexity :) but yeah it's definitely less intuitive and the purchase prices might make people go "huh?"

It would probably work best if the game logged the individual bids similar to how I demo'd above in order to be transparent about why people got the factions they did for the price they did.
If the text at the start is something like "Select the maximum price you would pay for each faction", and then the game log at the bottom showed bid-by-bid how it played out, I think that would be pretty clear. (And obviously, there could be a page explaining how it works, like there already is for the existing auction)

(The downside of needing to log step-by-step would be you couldn't take shortcuts in the algorithm, but I don't think it would be too much of a burden on the sever. There's the chance people could deliberately / coincidentally place matching very high bids which could add a few hundred "moves" to work through but as long as you batch them for writing to the db that shouldn't slow things down)

[AntoineBR]

I think the official app for Gaia Project will go with a closed bid system, there is an ongoing thread about it but the developer on BGG.