Fame mechanism

[kyrub]

Diagnosis
Fame is one of the most original concepts in 4x games. Yet its mechanism is somewhat disappointing: it triggers in the early game (heroes) when you have not built any substantial fame yet. Then you face a long "plateau" with very little activity. You gather fame, but the benefits are pretty poor. Being famous helps a bit. A little bit.

Formula

Code:

[b]p (item or hero) = fame / 25  + 1[/b] compared with d100
(for famous, p1 = p * 2)

So, when you hit about 20-25 fame (which is probably near endgame), the mechanism finally starts to work again. You are offered items, heroes, mercenaries... sadly, you don't need them that much at that point.

Let's not make mistake, the fame mechanism needs to be very hard to predict, unforgiving. There are other "primary" ways to get top units, special items, best heroes. Increasing the fame effect could be fatal to other concepts.

Suggestion
Let's do nothing but soften the scale.
We can use 1/2 percent steps instead of 1 percent steps, which gives us:

Code:

[b]p (item or hero)  = fame / 12 + 2[/b] compared to d200

Now, when you hit fame = 12, it makes a difference! You have 1/2% better chance (or full 1% in the case you're famous) of a hero knocking on your door. At fame = 25, you have the same chance (+1%) as in MoM 1.31.

With mercenaries, it is even better. Original formula

Code:

[b]p (mercenary) = fame / 20  + 1[/b] compared with d100
(for famous, p1 = p * 2)

Softened one

Code:

[b]p (mercenary) = fame / 10  + 2[/b] compared with d200

gives you a nice side-effect. Being famous (+10 fame) instantly brings you more attention from mercenaries. Instead of 2% chance (MoM 1.31) you get 1,5 * 2 = 3% chance of being offered help. Just cause gives you +0,5% chance as well. (These small % increments are quite important in long term.)

Summa of effects of softened formula:
- Fame is not useless in mid-game
- Famous trait can now be important beyond first hero offer
- Mercenaries will show up more often
- Items will be offered earlier than in endgame

(By the way, the increments could be even smaller, for instance 0,25%. With this type of turn-to-turn probability, even these small differences matter.)
What do you think?

[Catwalk]

So you're saying it only works on chunks of 25 fame currently? No difference whether you have 0 or 15 fame, other than for purposes of unlocking which heroes can come? That's pretty bad, I concur with the proposed fix. I'd also be happy with softening it even more, all the down to 1 fame point if that doesn't cause any problems. How about making current heroes affect recruiting chances even more? Or rather, increase the chances of recruiting heroes when you only have one or no heroes. It sucks to go for a long time without any heroes offered while you have cash piled and waiting for them.

[kyrub]

That's pretty bad, I concur with the proposed fix. I'd also be happy with softening it even more, all the down to 1 fame point if that doesn't cause any problems.

I am quite unsure about MoM in-game RNG for big numbers, it should be called suspect at the very least. Note, that the game itself, when going for a task of choosing from a list of weighed possibilities, first sums up those number weighs, and then it divides them until their sum is about 500. Only then, it performs the d500. Simply said, the game routines were programmed to avoid the big number RNG.

I think we can use this d500 as the benchmark: if the game uses, we are probably safe to use it 3 times in a row in the Random offer subroutine (heroes, items, mercenaries). This gives us this result:
Mercenaries (1 increase per 4 fame)
Items (1 increase per 5 fame)
Heroes (1 increase per 5 fame).

That's surely not perfect, but probably as good as we can get without breaking anything. And 5 times better than original!

Still - if you wish - we can push these numbers down. I could see 1 in 3 for heroes (it is divided later, see below, so it has smaller effect after the 1st hero) and 1 in 2 (or 1 in 3) for mercenaries. Although mercenaries' offers can become annoying later in the game with high fame (like every 10 turns), so maybe 1 in 3 is better. Items, I don't know. Every increase gives you extra 0.2% chance, less for heroes after the first one.

How about making current heroes affect recruiting chances even more? Or rather, increase the chances of recruiting heroes when you only have one or no heroes. It sucks to go for a long time without any heroes offered while you have cash piled and waiting for them.

The game already uses a 2nd mechanism (heroes only), which highly favours the situation without any hero. After that, its impact is deadly.

Code:

1. p1 = fame / 25  + 3
2. (for famous, p1 = p1 * 2)
3. p2 (hero) = p1 / [1 + (1+Nr of heroes)/2]
4. p2 is compared with d100

(all divisions are rounded down)

So, after the first hero (chance 3%), your chance drops by 50% (to 1 %), after the third by 66% (1%), fifth 75% (0%)... Just softening will give us a better result for after-1-hero (1,4%) and after-5-heroes (0,6%). With every +5 fame potentially increasing the chance.

Again: I know these numbers sound funny little, but in long span, they make big differences.


If you are not satisfied with this harsh second part of equation (I am not), I'd suggest another softening:

Code:

3. p2 (hero) = 2 * p1 / (2 + Nr of heroes)

.. which gives you:

Code:

0 heros .... div by 1   (1)
1 hero ..... div by 1,5 (2)
2 heroes .. div by 2    (2)
3 heroes .. div by 2,5  (3)
4 heroes .. div by 3    (3)
5 heroes .. div by 3,5  (4)

(in parantheses are the vanilla dividors)

Much better, especially for situation with 1 hero.

[Catwalk]

I think we can use this d500 as the benchmark: if the game uses, we are probably safe to use it 3 times in a row in the Random offer subroutine (heroes, items, mercenaries). This gives us this result:
Mercenaries (1 increase per 4 fame)
Items (1 increase per 5 fame)
Heroes (1 increase per 5 fame).

This would be acceptable, thanks. I do like 3 and 2 better, up to you. I rather like frequent mercenary offers personally, don't bother me any.

Code:

0 heros .... div by 1   (1)
1 hero ..... div by 1,5 (2)
2 heroes .. div by 2    (2)
3 heroes .. div by 2,5  (3)
4 heroes .. div by 3    (3)
5 heroes .. div by 3,5  (4)

(in parantheses are the vanilla dividors)

I kind of like a much steeper curve, giving much higher chances at 0 and 1 hero and drastically reduced chances further down. Mainly I want higher chances at 0 heroes.

[kyrub]

I like your approach. I was actually making attemtps in the same direction. It is refreshing to see us on the same side of the barricade, hehe.

First, an important analysis - how small % works:
Below is a table with three numbers. The first one means "how many turns to get equal chance of not-having-hero and of having-hero" (50%). The second one, "how many turns to have 90% chance of hero offer". The third one is a special "frustration barometer". If you play more MoM games, how often will you have no hero offer in 100 turns? As an example, in current setup (3%), every 21st MoM game leaves without any hero offer in 100 turns.

Code:

turns to 50%, turns to 90%, ratio of "frustrating games"

[b]1%[/b]
69, 228, 1/3

[b]2%[/b]
35, 114, 1/8

[b]3%[/b] - current normal
23, 75, 1/21

[b]4% [/b]
17, 56,  1/60

[b]5%[/b]
13, 45, 1/168

[b]6%[/b] - current famous
11, 37, 1/486

...
[b]10%[/b] - max
7, 21, 1/37486

Okay. Now, before finding the right formula, one has to ask himself:
-What percentage development do we want for 1st hero?
- And what should be the second hero like?
- Should there be a good chance for 3rd at all?
- And - important - how will it keep balance with Famous trait?

My responses:
1) 4-5% with high influence of fame. Maybe even, for every fame point, +0,2%. Which means, when you're frustrated, get more fame!
2) I would love to keep the second hero non-automatic. Rather rare. 2% seems about right, not more, obviously can be improved with fame. The game is strategically interesting when you have but one hero.
3) The third one, about 1% + fame.
4) Famous is the problem. Seirously. Because it doubles the chance and adds 10 fame - which we will make much more influential. Ideal for me, first hero, 10% with a good shout at second for 5-6%, third possible with about 3%, fourth unlikely (1,5%). Still, it may be too strong. We could give it less fame, but then it fails to attract better heroes... I don't know.

EDIT:
And this is the mathematical model for my suggestion. Works well for all I said, including Famous.

Code:

22 + fame - 5 * heroes
------------------------------
1 + heroes

compared to d500 (every point = 0,2%)

• 4,4% chance for first hero, 1,6% for second, 0,2% third.
  
• Famous: 10% first hero, 5,4% second, 2,8% third, 1,6% fourth.
  
• every point of fame counts.
  
• very little chance to get a hero with 4 already at your disposal.
  

[Asfex]

1.the 0 fame chance to hire hero must be same as vanilla.
2.Famous must double chance to meet hero,undepend of number of hired hero.
3.effect must be start with step of 10 fame or less. This mean famous can have 3x chances than no famous.
4. At the high fame(200) the result must be same as with famous.