Ok so to start this of:

*We have an issue: The tech market variations break players in the sense that it's 'wrong strat; gg no re' all sets now. These variations from set to set make 20m+ NW difference up and down which is game breaking. If s or Molotov were forced to war last set the market would have been way worse as they were the only 2 sentient techers.
The 'wrong strat; gg no re'-state is unacceptable to me in the long run. I have picked the wrong strat 3 sets in a row now partly as an experiment but also to try and find out if there are ways players can adapt to it, and my conclusion more or less is that we can't because the trend is shown too late in the round(other than a techer quitting if tech prices are too low, which is very good for the game).

*This is a BRAND NEW issue: The old way the tech buying bots bought tech, however flawed as it was, had one great quality; they spent more on tech the lower the tech prices were and less the higher the tech prices were. This was an unintended consequence of spending everything they have on tech, because cheaper tech => more tech => more income => more money to buy tech to a point.
This is no longer the case. Now cheaper tech means yes they buy more tech but once they reach a certain level of tech they value techs even lower than the market price, and so they refuse to buy it. This point comes sooner the lower the tech prices are. So we have lost the feedback effect that lower prices means bots spend more on tech.

------------------


Now with that said how can this be fixed in a sensible way?

1. I have long suggested rainbow bots "switch" into whatever strategy there is a noticeable lack off, according to prevailing market trends. This would soften the blow in my thinking. This might be hard to do and might be too slow to create enough effect? Could be enough that if tech prices are super high on friday and saturday they wont be on Sunday.

2. I have suggested also to change the tech value in the buying function so that bots will buy slightly more tech if tech prices go lower and slightly less tech if tech prices go high. This i believe is quite easy to do potentially but might not give enough effect on its own but should be customizeable.

3. The third thing I have suggested is that techer bots start exploring again when the tech prices go under say 1000+turnsleft*1.5, thereby adding land to the game instead of tech. this would be rather easy to do as well. This could be tuned between teching 20% of their turns to teching 90% of their turns depending on how high and how low the tech prices are.

4. Techer bots could be forced to sell tech at autobuy($1000) if average tech prices are below a certain number and conversely be forced not to sell at autobuy even for low priced techs if the volume weighted average price for all techs is higher than a certain amount. this would cause more of these techs to be on sale at lower prices which would naturally make techer bots tech more of the high priced techs.

I'm not sure which of these would be enough to accomplish a stabilizing feedback loop. I believe all should be done eventually but #1 seems like more distant and perhaps not as well suited to solve this current problem.

Anything players can do? Any other suggestion for how to make the bots respond to close supply gluts rather than make them worse?

#1 this is why the new leaderboard should have a 'best of strat' tracker, so even if you went casher you can still grab a point for the leaderboards if you have the best casher
#2 This past set was a 'gg no re' to you? but uhh, your commie scored 72m and won... maybe it was 'gg no re' for a casher, and farmer still doesn't seem that great

Yea I mean erm... a commie won. lol And the other big commie was really close too.

So I'd actually say, as the only person who's played a large farmer under the new rules, at the same acreage I outproduce theo casher as a fasc farmer (by a pretty large margin). I'm not sure about rep casher, but the R is prob a big target. lol

There are certain adaptations to the 180 extra turns and the rest of the changes that has not been done yet by most players, which factors into this.

Casher and farmer could have both finished a lot better than they did because so far no one seems to know how to play them yet since no one played them seriously before these changes made them viable. Indy is the same as it has been forever in express just its easier to get 50k+ land with X turns to go now since we have more turns and better grabbing. But it doesn't change the fact that this was a techer reset. Techer has also changed and players have not adapted yet.
The reason commie indy can succeed even with only 2 sentient techers is because it's good in general on express and it is also the strat that gains the least of it's income from tech. That said I lost 20-30m NW on picking the wrong strat both last set and 2 sets ago and it's undeniable that the swings from set to set are far bigger than what we have seen for the past 200 or so sets and that last set was worse than normal for all tech buying strats:

https://i.imgur.com/1gBYXne.png

You can go back further if you want but when the tech market last was this high we had 180 less turns, casher and farmer was non-competetive, grabbing returns were lower because of fewer bots plus the ghost acre limitation that was removed and the top 10 had 6-10 techers usually as a result, so it wasn't the same game.

Either way the swings are bigger, I have provided an explaination for why the swings are bigger, I have suggested they are a problem. You are suggesting they are not. I think because you have not experienced being on the wrong strat yet.

Not only me. Rob quits every set that he picks the wrong strat so he stabilizes the market to a degree. :)

Hey thats not entirely true - i did play on last set & just overslept at the finish.

But largely, yes thats true. I'd like to claim i do it for the sake of the server. You're welcome people :)

I do however think the best way to fix this is seasonal leaderboards with best of strat points.
Then the incentive to maximize on a wrong strat is still there & keeps it interesting regardless of what the market is like.

Its hard to keep motivated when you're on a wrong strat and you have nothing to play for.

When you say it like that, seasonal leaderboards exacerbates the initial problem by making players ignore it, by changing the purpose of the game from "winning" to "winning against those of the same strategy as your own".
I do see the merit of it. It's like when I participated in the 0cs or the tyr casher challenges years ago, I was competing in an own tranche and was happy with that. Winning and total NW wont be meaningless in seasonal leaderboards tho.

Also Im not sure if it changed meaning somewhere or if there is some misunderstanding either of us have of it but this is the proposed seasonal leaderboard metrics:


That means if Tmac played a casher to say 70m NW this season and I choose casher another reset later and that is a horrible casher reset, I will not be playing for anything but total NW(which may or may not be enough to keep me going).

Because this is a problem that has been created by the very recent bot changes(that have been good for the game in general) I believe bot tweaks should be the first line of dampening this trend. Current objections to this view seems to be entirely based on a, from my understanding, flawed idea of what seasonal leaderboards will do.

Basically molotov and rob in this thread assumes that Molotovs suggestion for how the best of strat thingy will work is the one that will be implemented and not the current proposal which I quoted above. If that is the premise I somewhat agree that this solves the issue since we have something else to play for even when the market is horrible for us.

Molotov's post on this here:

Personally, I favor slagpits initial plan better, which would NOT solve this issue. The reason I like it better is because otherwise its almost impossible to beat someone who specializes on one strat only, on that best of strat metric.

5. The Casher, Farmer and Indy bots could dump income techs on the market AFTER their destocking has begun IF the market prices allow for it to sell at above say 1200+Turnsleft*4. The price they sell at will be modelled after one of the techers pricing algorithms that suits the purpose. Or could possibly even set the sale price as a static constant, saying for $1800 you may buy any amount of tech you want (this will be more tech than any number of humans can buy).

This would mean nothing for techers, but that there would be a ceiling tech price in the last 2-5 hours of the set which allows the non-techers to max their techs if they want even if tech prices have been super high during the set.

“Best of strat” I think is better than those “best commie net” etc ones but i wasn’t trying to argue we shouldn’t have the other ones, just that BoS could be included on the leaderboard

Ohh so you mean that 4 players each set get a BoS award on top of that; the best indy, the best casher, the best farmer, the best techer. I can totally get behind that, that is a great suggestion.
But we still have extreme market fluctuations making all the other entries into the Seasonal leaderboard mostly about predicting which strat will be best as many sets as possible. Also, I'd like to point out that there are no seasonal leaderboards right now.

Can I be considered a farmer if I made a technical start for 800 or more moves while selling technology when it has a high cost, then I built farms. And then at the finish line made a revolution in theocrat?

I would call that a farmer with a rather long tech start, MILORD! :)

OK . I just clarified in order to understand to which category such a country should be assigned. This set, as I said, I will play as a democrat-farmer, without a technical start. I will be testing this combination shortly. For a complete view, you need to play several sets. In the future, I plan to test a fascist farmer, a kesher republic and a kesher theocrat, a tyrant farmer with a technical start. Kommunist industrial does not inspire confidence in me, on this server.