PCGamer: Soren Johnson says Civ 3's bargaining table was a 'big mistake' (GDC22)

[tjs282]

It also appears to work for gpt, though gpt deals have some inefficiencies in many cases.

AFAIK, the only real inefficiency of a GPT-deal is the one that Firaxis deliberately programmed, i.e. the [gold-]total acceptable to the AI will always work out higher for a 20-turn GPT-deal — by about 10–15%, as far as I can tell (rough estimate!) — than for a one-time up-front payment. But provided that your own towns/economy continue to grow over that 20 turns, those "interest-payments" aren't generally problematic (especially if you manage to goad the AI into breaking the deal early!)

In my example, I ring up the Maya and offer them Philosophy for Mathematics and The Wheel. Smoke will probably like it, so I then click lump sum and I get an autofill of 29. That's the exact optimum number for that deal.

And if you took off the lump-sum and clicked "Per turn" instead, he'll accept 2 GPT (40 gold total), because 1 GPT won't give him the 29(+) that he wants. But if you offer him 1 GPT and then click "Lump sum", his up-front payment-request will likely go down to 9–10 gold.

That's one of the things I really like about this patch: based on what I have in my Treasury, and how much income I'm getting, I can very quickly zero in on an optimal deal for my current situation.

 

[Spoonwood]

AFAIK, the only real inefficiency of a GPT-deal is the one that Firaxis deliberately programmed, i.e. the [gold-]total acceptable to the AI will always work out higher for a 20-turn GPT-deal — by about 10–15%, as far as I can tell (rough estimate!) — than for a one-time up-front payment.

I was referring to cases like where you would pay 2 gpt in your example instead of 1 gpt and 11 gold. It's 18 gold for 1 gpt.

Also, the other issue with gpt is that I can't always track where it went. On a pangea map with scouting it's possible to purchase a tech early with gpt, given you can get enough gpt. From my previous read, it can disappear. It's easiest to test with mass regicide enabled with the king units used as scouts and minimum aggression.

If I recall correctly, I've done deals like let's say the Iroquois have Construction and Arabia has Polytheism. I first talk to the Iroquois and purchase Construction with gpt, suppose it's 20 gpt (not a projected number for any game; I think I used the change all citizens to tax collectors trick once or twice). Sell Construction to Arabia for their lump sum and Polytheism. Come back to the Iroquois. Haggle about what they will give. They will take Polytheism giving 6 gpt. They won't take Polytheism for 7 gpt. If I remember correctly, the 7 gpt can be different than the "we're getting close to this deal", which if I recall correctly, is the one where they have more gpt, but the value you're asking for is higher than the value of the tech. But in cases like this they say something like "they will never take this deal". They don't have more than 7 gpt even though you gave them 20 gpt, suggesting that they were in a deficit, or they have some amount of gpt that they have to produce before they will give you gpt. You can't get that gpt back if it's not in the economy on that turn at least. I don't think that the Iroquois in such a situation would necessarily start getting lump sums of gold every turn. I think they would still have no more added gold before markets get completed in the next few turns if you gave them 4 gpt. My read is that such deals can result in the gpt disappearing from the world's economy. I might be wrong about that. At the very least, the gpt has disappeared from my sight, unlike how I could still tell that a lump sum of gold would exist. I didn't check to see what would happen with such negative gpt n-fers over time.