Strength tweak proposal

[some12fat2move]

I propose that the way STR affects armour EVP be changed to aevp=(3evp^2)/str.

Here is a graph to help visualize it. The first line represents fire/ice dragon armour, the second line represents plate armour, and the third line represents gold dragon armour. The x-axis is your strength and the y-axis is your AEVP.

The graph evidences a couple of important points of this proposed formula:
-If your strength is triple your armour's EVP, you get the base EVP of the armour, which is what happens currently.
-If your strength is double the amount needed for base EVP, you get half EVP. And if your strength is quadruple what's needed for base EVP, you get quartered EVP, and so on. Conversely, if your strength is half the amount needed for base EVP, you get doubled EVP, and so on.

This will affect gameplay because:
-This proposal makes strength above your armour's breakpoint still help you reduce EVP, making strength very useful for anyone wearing cumbersome armour even if he/she has already exceeded the breakpoint. The gains from strength will be much greater in heavier armour.
-This proposal eliminates the hard breakpoints involved in the current strength and evp system.

Thoughts?

[Galefury]

This is the same graph including the current aevp for comparison. So this would not only lower the armor penalty when you have more strength than the current breakpoint, it would lower it for any strength value greater than the base penalty (a third of the current breakpoint). I don't know if that matters, and this change would at least get rid of the breakpoint. The formula aevp=(3.5*evp^2)/str would approach the current curves more closely.

BTW, this makes gaining strength less important for strength values slightly below the current armor penalty breakpoint by reducing the gain from each point of strength. It makes it more important above the current breakpoint by introducing a gain from strength where there currently is none. It makes it far more important than it currently is far below the breakpoint.

[galehar]

I really like the idea, I have been thinking about making such a proposal. Thanks for coming up with the formula and numbers
I also like that it keeps the point of reference that str=3*EVP means that you get base EVP, even if EVP is a bit abstract.

The formula aevp=(3.5*evp^2)/str would approach the current curves more closely.

I think it'd be better to adjust the individual armour stats to adapt to the new system. Let's change the magical 3 if we decide that the slope of the curve needs fixing, but it looks like a good start.

[some12fat2move]

Well it's been a bit over 24 hours since the last response and the proposal looks pretty good. Galefury did an excellent breakdown of the differences between the current system and the proposed and nothing objectionable was found except the danger of power creep, which can easily be remedied later by changing the multiplier, or individual armours, or both. Minmay pointed out that the str/dex situation could be a lot better but agreed that the proposal is an improvement. So with the devs' blessing, this proposal can move to the next step. I know things aren't implemented straight from the tavern so is there somewhere else I should post this?

[nicolae]

Well it's been a bit over 24 hours since the last response and the proposal looks pretty good. Galefury did an excellent breakdown of the differences between the current system and the proposed and nothing objectionable was found except the danger of power creep, which can easily be remedied later by changing the multiplier, or individual armours, or both. Minmay pointed out that the str/dex situation could be a lot better but agreed that the proposal is an improvement. So with the devs' blessing, this proposal can move to the next step. I know things aren't implemented straight from the tavern so is there somewhere else I should post this?

Probably the next step is to discuss it more in the ##crawl-dev IRC room, and then submit it as a patch to Mantis.

[some12fat2move]

Probably the next step is to discuss it more in the ##crawl-dev IRC room, and then submit it as a patch to Mantis.

Ugh, I'm not fond of IRC, but I guess I can get around to it sometime next week.

[Galefury]

Or you could just submit a patch.

[galehar]

Probably the next step is to discuss it more in the ##crawl-dev IRC room, and then submit it as a patch to Mantis.

Or you could just submit a patch.

There's almost no coding involved, so the next step is indeed to bring this discussion to ##crawl-dev to get other devs' approval.

[Galefury]

This usually leads to someone joining ##crawl-dev, making a suggestion, then leaving five minutes later when he doesn't get an immediate response. If a patch is put on mantis, even if its a one line change, there is something to talk about that doesn't just go away after five minutes.

[Bomanz]

Why not just have each point of STR give 2 HP up to 20, and 1 HP for each point thereafter? I think the damage bonus should be increased as well. Would like to see STR more valuable for non-melee types.

I think that DEX would be the new weak stat after that though...

[Bomanz]

I suggested a while ago to make strength give AC (this is better than HP in my opinion - stats jump around a lot and max HP probably shouldn't).

AC could work out pretty well. Although in my experience STR is not a huge target of stat drain usually, and there already is ROT in the game. Why shouldn't MAX hp fluxuate a bit with equipping for example a 5 str piece of armor that may bring you up 5 points?

[BlackSheep]

Say you have a +5 strength ring. You put it on, your max HP goes up. You take some damage, then take it off. Your max HP goes down, but what about your current HP? If it stays where it is, you're going to invite people fiddling with their equipment to "maximize" the amount of damage they can take. If it goes down along with max HP, you can have bizarre scenarios where you take off your ring of strength and keel over dead. We could introduce code to check for a bunch of corner cases, or we could just have strength raise AC which in 99% of cases is the same thing as increasing HP.

[Bomanz]

Just make it so you can't die when taking off str items, instead just staying at 1 HP? Lots of effects are like this.

Anyway, AC would be fine and easier.

[galehar]

Here is an update to the proposal:

5*evp^2/(str+2) but with some adjustments. See the graph. Brown is leather and WDA (-1 and -2), not much changed. Green is fire (and ice and pearl) dragon armour (-3), penalty is a bit increased which is probably good. Light blue is chain, splint and plate. Chain stays at -4 but since it is a nerf, it gets +1 AC to compensate (8). Splint (-5) is removed, and plate takes its place at -5 (little buff). Not sure if SDA needs to be buffed to -4 or nerfed to -5. In yellow are CPM and GDA. GDA gets buffed to -7. CPM is buffed to -6 to have a continuous range, maybe lose 1 or 2 AC to compensate.

[Lasty]

If I understand this graph right, It looks like the -7 armor (GDA) would have an effective penalty of a little over -10 at 27 strength on this plot. The -6 armor (CPA) would have an effective penalty of -9 at 24 strength. It seems like these changes makes both of those armors worse, even for high STR characters.

[galehar]

If I understand this graph right, It looks like the -7 armor (GDA) would have an effective penalty of a little over -10 at 27 strength on this plot. The -6 armor (CPA) would have an effective penalty of -9 at 24 strength. It seems like these changes makes both of those armors worse, even for high STR characters.

Nope, you're misreading. Both GDA and CPA are buffed at any str value. Their new curve (orange) is completely under the old one (yellow). GDA is at -8.16 at str 27, CPA is at -6.66 at str 24.

[nicolae]

(this would also be a great time to add iron/shadow/quicksilver dragon armour!!!!)

What effects would those have?

[mattlistener]

Say you have a +5 strength ring. You put it on, your max HP goes up. You take some damage, then take it off. Your max HP goes down, but what about your current HP? If it stays where it is, you're going to invite people fiddling with their equipment to "maximize" the amount of damage they can take. If it goes down along with max HP, you can have bizarre scenarios where you take off your ring of strength and keel over dead.

An answer that avoids weirdness and coding-for-corner-cases would to have "current % of max HP" remain the same when an equipment change alters Max HP.

So: you currently have 15/20 health (and no regen). You make an equipment change that gives you +8 HP, and you go to 21/28 health (same ratio as 15/20). You undo that equipment change, and you go back to 15/20.

[Fergy]

An answer that avoids weirdness and coding-for-corner-cases would to have "current % of max HP" remain the same when an equipment change alters Max HP.

Wouldn't this also cause problems with equipment switching for the purpose of faster regeneration?

(btw, I think your idea would work if the +HP was only based upon base-STR, and not include STR bonuses from items)

[rebthor]

(this would also be a great time to add iron/shadow/quicksilver dragon armour!!!!)

What effects would those have?

High AC/ponderous for iron.
?? rTorm I guess for SDA (maybe 50% torment damage? no idea really)
Low AC but inherent swiftness for qda.

[galehar]

That probably happened because you linked to a misleading graph. It has a line for -7 EVP and a line for -8 EVP, none for a -6 EVP, even though nothing will have an EVP of -8 under the proposal.

Indeed. Sorry about that, it was late when I finished this. Here is the link to the good graph.

[galehar]

I've made another update to the proposal.

(5/2) * evp^2 / (str+3)
The EVP has a bigger scale about *3 (GDA is 27). The penalty to dodging bonus is adjusted to use EVP/3 instead of EVP. The bigger scale should make thing easier by giving us more room to balance and differentiate armours.

[galehar]

A better scale is not defined by being coarser. When they are too fine, we make them coarser, but if we have issues balancing because of a coarse scale, why not make it finer? While balancing the new formula, it appeared that it was too coarse. Some armours needed 6.5 or 7.5 EV, so I suggested this move to avoid decimals.
Regarding developer support, for now at least elliptic supports it. We'll see about the others.

Edit: oh, I guess you meant support for reverting the aptitude change. It was a good change, I have no intention of reverting it. The aptitude scale was too fine and we made it coarser. The EVP scale is too coarse, I propose to make it finer. I'm not sure why you think the comparison is relevant.