Quote:"Accept trade if the TradeValue(OwnSpell) >= TradeValue(OfferedSpell)"
Yes, in this way, both the AI and player could "trade up", but only if brokering was allowed. If it wasn't, they still can't.
That's literally the same thing I just said, that rule allows the AI to broker spells upwards.
That + "trade brokering is allowed" -> it's impossible to do this because AI would trade up.
That + "trade brokering is not allowed" -> It's perfectly fine.
We have to keep that in mind as a consequence after we make a decision on trade brokering, but it's neither a proof or argument for either side. It just shows your suggestion to make AIs maximize trading requires to ban trade brokering first.
As we don't yet have conclusive evidence on trade brokering, nor a decision, and this depends on it, we can't do it in the current state.
...maybe there is a different solution though.
We could have this rule instead :
"The AI players must trade spells with a value difference of 7 or less and neither AI player can receive a spell with higher trade value than the highest trade value spell they already researched"
Which does prevent trading up, with or without brokering, unless they get lucky and find a higher tier spell in a lair but that requires beating that lair which sounds unlikely for an AI that's a whole tier or two behind others.
That said, I'm still not sure more AI trading really is what we want in the game. I mean, "The human player gets too much spells from trading" and "The AI players don't get enough spells from trading" are the two sides of the same coin but ultimately there is the edge of the coin - the intended pacing of the game. So it's very well possible neither side, or both sides need adjustments and it's possible they need changes in the same, or the opposite direction.
So before we can do anything at all, we need to pretty much see from testing which if any are true :
-Players can have higher tier spells than the intended pacing -> We need to do something with human trade
-Players can fill in the gaps in common/uncommon spells in their spellbook by trading (you did report this) -> intended feature, not a bug
-Players can fill in the gaps in rare/very rare spells in their spellbook by trading -> probably the result of AIs not dying fast enough, if not then we need to do something with human trade
-AIs can have higher tier spells than the intended pacing -> We need to reduce AI trading
-Players or AIs are a tier behind the intended pacing -> This can't happen without a major bug because trading wasn't part of economy calculations so there should be enough resources to make it impossible.
-AIs cannot fill in the gaps in common/uncommon spells in their spellbook by trading -> We need to make AIs trade more
-AIs can fill in the gaps in rare/very rare spells in their spellbook by trading -> AIs should have killed each other before this happens, but even if they don't, this is probably fine as it makes the game more challenging. Dual realm AIs tend to be easier as is.
-AIs cannot fill in the gaps in rare/very rare spells in their spellbook by trading at all -> This is the risk of playing multiple realms and the AI is protected from it by having a "guaranteed reroll" feature to get at least one good spell or spell combo. Not a problem although it's not desirable either.
So before we can really make any progress in any direction, we need test data for all the above except the second row and preferably from multiple people.
Quote:You can't trade 5-10 different players, realistically, even in a 13-Wizard game, and even if that many trading partners were available,
You did say you traded Invisibility to 5 different people for 5 different uncommons. You could have found that Invisibility in a lair or could have researched it yourself. Yes, you didn't, you traded for it but that still means two instances of brokering and four instances of duplicate trade. Unless you automatically count all duplicate trades as brokering as well which you shouldn't because then you won't be able to measure how much the brokering affects the system without duplicate trading, which is what we need to compare the two.
As for how many players you can trade a spell to, that's actually possible to calculate. Assuming you are trading away leading tier spells which no one has yet and have high enough trade value to trade for most other spells in the same realm, which is the worst case scenario, we get :
1 realm wizard : There is one mono and 4 dual AI templates including your realm out of 15, so the chance is 1/3 per AI player, or 13/3 = 4.33 players in a maximal player count game.
2 realm wizard : There are two dual templates, one that has both your realms, 3 that has one of your realms and 3 that has the other one of your realms, for a total of 9 out of 15 templates, which equals 60%. In a 13 player game that is a 13*6/10 = 7.8 players who has at least one realm with you. This doesn't matter though because we assumed we trade one specific spell and one spell can't be both realms at once. So the average number of trading you can do per spell you own is 4.33 if you don't adapt to circumstances. If you do and pursue the research of a spell in the realm with more trading partners, you will most likely see 5-7 players you can trade the spell to on average.
In either case, your rare spell with value 30 will trade for
at least 4-5 other rare spells that have value 30 or lower.
Compared to this, trade brokering only ever allows trading spells down. While it's theoretically possible to always trade for only exactly 1 value lower, it's not realistic. In general, you'll be trading for about 4-5 value lower each time at least, both because the exact trade values are not shown to players but also because the spell with the perfect value won't always be available on your first try and you can't afford not trading a "still pretty good" offer. Assuming a drop of 4-5 value each trade, a rare spell with value 30 will only trade for
2-3 uncommons (of values approximately 15, 20, 25) and
some commons (trade value 0-15)
So looking at just the raw numbers, duplicate trading gets you better spells than brokering : same quantity but higher tier spells.
This assumes the "at least +4 value" code gets removed, otherwise it's safe to say the average trades the player can make will trade down by at least 7-8 each time, for the same reason, it's extremely unlikely to nail the perfect trade where the received spell is the highest possible, without literally looking up spell values in tables.
Note that all of the above assume the trade has no modifiers on either side, so either both you and the AI get a spell you can research on your own, or neither of you do.
If you consistently trade away spells the AI can't research in exchange for spells you CAN research then you can keep up brokering as long as you want or get 4-5 rares slightly better than your offered spell but the value of that is questionable as it doesn't give you new spells at all.
Now if I assume you are trying to trade away something that's not "leading tech", like a 25 value rare when most people have 30 value rares, it actually doesn't change anything. You can still trade for 5 spells of equal value and it's actually much more likely you can trade successfully because the AI prefers to research the good spells first, so if you did go after the "leading spell" but wasn't the first to get it, your chances to trade it away are drastically worse. If you picked a mediocre spell the AI doesn't prioritize much then you are almost guaranteed to be able to trade it to everyone, even if you lose the research race because no one will pick that thing until they got all the better stuff first.
In case of brokering however, this definitely makes the results worse as the starting value is lower so the chain contains one fewer spells before value shrinks to "untradeable garbage common".
All of this assumed we don't mix brokering and duplicate trades and only use one or the other, because we want to measure their effects separately.
The theory seems to show duplicate trades are more beneficial, but the effect of luck is higher on brokering. (Do you trade for 1 or 7 value lower consistently? You can't know and even if you do the trading partner might only have the 7 lower value spell, but it is the main determining factor for how much value you get out of brokering.)
Either way we need more test data, if there is not problem to fix, it doesn't matter which contributes more to gaining spells, duplicates or brokering.
