Hello, some notes: 1) yes, AI can use recalls if a recall list is provided. This, however, changes little wrt balance.
2) Also, it'd be a good thing to allow the recall cost to be calculable by WML. 3) an alternate idea to be allow the player to recall up to N units (i.e. 'recall only during the first turn(s)' or 'recall up to N units'), regardless of available gold. This way, the player can pick some upgraded units from the previous scenarios as 'elite', but would be forced to recruit new units afterwards. On Thu, Jan 26, 2012 at 12:26 AM, Dan Gerhards <[email protected]> wrote: > This is well reasoned and argued. I could get behind this, FPI or not. > There are a couple problems I see, but I have a suggestion for fixing > them too. > > First, I am not in favor of using 100% of recruit costs. As you say, it > would change Wesnoth from a game where you build an army to a game where > you manage finances. That would rob the campaigns of much of their > appeal (for me at least). It would be easier to test, but that's not > worth the other problems it would cause. So, I do think recalling should > be less than recruiting for high-level units. > > Second, recalling level-1 units should be *more* than recruiting them > because of their experience and known traits. > > I propose a system more like this: 10g + 50% of the recruit cost. This > would make most level-1 units a bit more to recall, and higher-level > units less. The higher they are, the higher the discount. > > To still make the gold of the AI more comparable to that of the player, > we should make (and use!) a mechanism that easily generates a recall > list for the AI leaders. This also has the advantage of making it more > interesting as you are fighting lots of units with experience. It could > be done pretty easily with a macro, but it would be better in the [side] > tag as "generate_recalls=(number of units)", or something similar. > > ---Dan Gerhards aka beetlenaut > > On Wed, 2012-01-25 at 22:43 +0100, Fabian Mueller wrote: >> Hello Devolopers, >> this is an issue that is on my mind since long but I didn't dare to talk >> about loudly because It concerns an old but unlisted FPI. >> >> The FPI in question is: >> Different recall costs depending on the level of the recalled unit. >> >> Most proposers go with the proposal for 20 gold per level. >> >> I want to present a discussion about that topic but let me introduce a >> vocab to argument with first. >> >> The "Downwards Spiral" >> >> I call a phenomena that can easily be experienced when playing Wesnoth, >> in fact most or even all Players will have encountered it at least once, >> the "Downwards Spiral". >> >> Consider you enter a scenario with a critical amount of gold. >> Very soon you discover that you need to spend all your gold to win the >> scenario and even worse >> that spending all gold with a balanced set of recruits/recall isn't working. >> >> Let's go in detail what a balanced recruit/recall set is and why it is >> not working. >> A set of units can be called balanced when it consists of more first and >> not maxleveled units than >> maxleveled ones. >> >> You need the first level (or zero level) ones to be sacrificed in place >> of high level ones in critical situations. >> The not maxleveled units are needed to suck up the experience points and >> the maxleveled to solve some situations where >> the enemy isn't listening to lesser arguments. >> >> The obvious problem with such a set is the lack of firepower. >> Compared to a maxlevel only recall you get much less bang for your bugs. >> Thus the level is still completable if you recall unbalanced but not if >> it is done right. >> >> So, what are the consequences of such a situation? >> It depends on how the player acts in the scenario. >> >> I think there are basically two ways to handle the problem: >> >> 1) You try to rush and finish the scenario early to gain a better >> financial situation, >> but that will lead to the loss of some or even many of your high level >> units. >> The fact that your maxlevel veterans afford a huge amount of upkeep to >> maintain can stand in the >> way with taking this solution, since you may end with not much benefit >> gold wise. >> >> 2) The player plays very cautious to prevent the loss of the maxleveled >> units. >> Retreading units often and taking no great risks will certainly work >> when done well but that takes time and >> there is no improvement of the financial situation to expect. >> >> Thus the next scenario is entered either with a insignificant amount of >> gold again or with a thinned out army to rely on or even both. >> >> One can guess that this may get worse from scenario to scenario. >> Finally the player realises that he has lost the game, not in the >> present scenario but some time ago in the past. >> He/she needs to restart an earlier scenario that maybe 2, 3 or even more >> scenarios ago. >> >> This is the point where many people quit playing the campaign or Wesnoth >> at all. >> Some of them complain that Wesnoth is to hard and totally unbalanced but >> their voice isn't recognised much. >> They are told that Wesnoth is designed like it is for a good reason and >> to learn how to prevent the mess early on. >> >> But I call that a design an epic fail (sorry, Dave). >> A good design is to let the player recognise the failure early on to >> keep him motivated. >> >> Now, let's talk about balancing. >> Balancing a campaign is hard because when you change something in >> scenario 3 you need to play all your way through 17 scenarios to >> playtest the 20th. >> This is not only an issue of gold but more of that there is no suitable >> recall list available. >> >> The current gold carryover system can be used in a way to give the >> player at least enough gold to win the scenario even with no carryover. >> And it is also a balancing paradigm that a scenario should be winnable >> by a good player in easy mode without a recall list with just the >> minimum amount of gold the scenario guarantees. >> Still this way of balancing does not lift the need to play through the >> campaign at all difficult levels. >> >> Let's assume that the "Age of Heroes" would be balanced as good as >> "Default" or that it is accomplishable in a reasonable amount of time >> and effort, >> considering that the proposal will bring in more feedback about the matter. >> Second, recall costs are set to the same amount as recruiting the unit >> would cost. >> Of course the player still can recruit only lvl0 and lvl1 units, so that >> is just a theoretical matter. >> >> This changes the situation in several aspects. >> 1) The amount of gold between the different sides/teams is much better >> comparable. >> Currently the power the AI can buy with it's gold is a static value >> since it relies on recruits only. >> It has to be compared to the player's forces which heavily depends on >> the recall list. >> >> Both issues would be eliminated. >> While the recall list grants the player the access to some special >> features the bare power of the force isn't depending on the available >> recalls any more. >> Secondly, the amounts of gold are directly comparable between the sides. >> >> As a result the playtester is able to select a single scenario and test >> it at any difficult level without the need to rely on a valid savegame >> with recalls. >> >> 2) This will change the character of Wesnoth. >> Gathering a strong army will get less important in favour of a proper >> gold management during the campaign. >> >> 3) This will certainly unbalance all current campaigns. This point might >> not that important since the actual system of recall costs can be >> customisable thus only converted campaigns will use a "repaired" system. >> >> 4) Last but not least this prevents the downwards spiral from showing >> its ugly head that often. >> >> 5) This will most likely lead to a better balanced "Age of Heroes". >> >> A reasonable implementation could work like "recall_costs=20", which >> reflects the current behaviour and be the default at first. >> "recall_costs=70%" would be an example for a less radical approach that >> lessens the effects of all mentioned aspects to some degree. >> "recall_costs=100%" means that recall costs equal the recruit costs and >> reflects the assumption made above. >> >> I plan to implement the feature during the next cycle and convert at >> least LoW to the new system, most likely going with recall_costs between >> 60 and 80 % of the recruit costs. >> >> Regards, >> Fabian Müller aka Fendrin >> >> >> _______________________________________________ >> Wesnoth-dev mailing list >> [email protected] >> https://mail.gna.org/listinfo/wesnoth-dev > > > > _______________________________________________ > Wesnoth-dev mailing list > [email protected] > https://mail.gna.org/listinfo/wesnoth-dev -- Cheers, Iurii Chernyi _______________________________________________ Wesnoth-dev mailing list [email protected] https://mail.gna.org/listinfo/wesnoth-dev
