training formulas

[weirwood]

A couple of years ago, I found a website that listed all the correct formulas for trainers, but I haven't found those on the web since.

Since there are some significant differences when training combat and magic with various trainers, the list was really useful. Does anybody have a copy? Or can these values be gleaned from the code, somehow?

[Sir_Toejam]

go to gamefaqs.com, and grab one of the walkthroughs. the ones by dan simpson always list all the trainers, what they do, costs, and locations.

http://www.gamefaqs.com/search/index.ht ... platform=0

[weirwood]

I know Simpson's walkthrough, but he doesn't have the correct values either.

For example, if you train Int and Magic with Perrin, the correct formula is something like:

+2 Int, +(Int-Mag+2)/2 Magic

[Sir_Toejam]

then, none of the walkthroughs do.

*shrug*

[keldorn]

Ya I too never came across such a walkthrough. I know that the first time Spark trains with De Snel, his Combat increases by around 6-8 points, 2nd time its 4 points. Therefore I'll keep trying to get the optimal results.

[weirwood]

It wasn't a walkthrough, it was an Ultima fansite. and while I no longer know the specific formulas for each and every trainer, I remember figuring out back then that the only worthwhile trainers (for your companions, at least) were Sentri (Dex & Com) and Menion (Str & Com).

[Sir_Toejam]

yeah, you're right about that; sentri gives more than just the dex mentioned in the simpson faq.

I too almost exclusively used those two trainers as well.

the puzzlement continues in SI as well.

If you train with shazanna, you get far more combat points than one would expect, although this does seem to taper off after a few sessions.

in the end though, none of it really matters. combat is so damn easy in both games after you get your guys to around 20 in all stats, it hardly makes a difference what you do.

[keldorn]

Yes agree that Sentri/Menion are the 'best' trainers in BG but I do use De Snel to do a major boost for those with low Combat skills first (Spark/Jaana) before Sentri steps in.

[weirwood]

Well, if you train twice with Menion, your Combat skill should rise as much as when you train with De Snel. It costs 4 more skill points, but you'll also get +4 Strength. And since you can't train Strength on it's own, it makes little sense to visit a specific combat trainer.

[marzo]

Here it is, straight from usecode:

[edit]There are a couple inaccuracies in this post. I posted the corrections below.

* When you train Strength, every +1 from the trainer increases Strength and Health by +1.

* When you train Dexterity, every +1 from the trainer increases Dexterity by +1 and sets Combat to ((Dexterity + 1) * Combat + Dexterity - 1) / Dexterity, rounded down, where "Dexterity" refers to the value before the +1 increase.

* When you train Intelligence, every +1 from the trainer increases Intelligence by +1.

* When you train Combat, every +1 from the trainer sets Combat to (Dexterity + Combat +1)/2, rounded down.

* Training Magic works exactly like training Combat, but replacing "Combat" and "Dexterity" by "Magic" and "Intelligence", respectivelly.

For example, Bradman (in Yew, IIRC) increases Dexterity by +2 -- this means that your Combat will be increased by the above formula twice, increasing Dexterity by +1 twice (once in-between and once after the two increases in Combat).

As a rule of thumb, training Combat instead of Dexterity is better if your Dexterity is higher than your Combat. If your Combat is higher than your Dexterity, however (which shouldn't happen in the original), NEVER train your Combat -- train your Dexterity until they are equal.

Likewise, you will lose less if you try to keep Dexterity+Combat (or Intelligence+Magic) an odd number when training Combat (or Magic).

[keldorn]

I see, now I understand why Spark gets a great boost to his Combat when training with De Snel. My next question is since this formula is generic, that means if he trains with any other Combat trainer, eg Markus, the increment should be the same right? If so, what is the advantage given by those trainers listed as "extensive"? Can't remember exactly but think, they are:

Bradman (Dexterity)
Zaksam, De Snel (Combat)
Inforlem (Intelligence)
Perrin (Magic)

[marzo]

Just a small note: I realize I accidentally left off a couple important bits above (one of which is inexcusable and affects my analysis above):

Minor Point: You can't train Dexterity if your Dex is 30.

Major Point: When you train Combat or Mana, if the value given by the formula is used only if it is smaller than the controlling attribute; if it is equal to, or greater than, the controlling attribute, then Combat or Magic will simply be increased by +1 instead (and unless that would put it above 30). Thus, for example, if you have Dexterity <= Combat, each session of Combat training will increase it by 1. More info on this below.

My next question is since this formula is generic, that means if he trains with any other Combat trainer, eg Markus, the increment should be the same right?

Not always, no;but sometimes. With the intent of making the explanation easier, I will borrow a small part of Dan Simpson's walkthrough; specifically, the following part of his trainer table:

Code: Select all

  Name:      City:      Skills Increased:           Cost: (Skill/Cost)
  ¯¯¯¯¯¯¯    ¯¯¯¯¯¯     ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯           ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
  Bradman  - Yew      - Dex +2                    - 30 gold (15)
  De Snel  - Jhelom   - Combat +2                 - 40 gold (20)
  Markus   - Trinsic  - Combat +1                 - 20 gold (20)
  Zella    - Britain  - Dex +1, Combat +1         - 45 gold (22.5)
  Sentri   - Britain  - Dex +1                    - 30 gold (30)

While the above table isn't entirely accurate (as it neglects the effects of the above formulas), it is correct in this -- each of the stats is incremented by the formulas above a number of times equal to what is indicated in the table.

From the table, you can see that Markus is rated as a Combat +1, while De Snel is rated as Combat +2. This means that if you train with Markus twice in a row, you will get the exact same benefit as training with De Snel once. Same thing with Sentri and Bradman. But things will *not* be the same if you mix training Dex and Combat; for example, training with Sentri and Markus is equal to training with Zella only if you train with Sentri first; if your Dex is less than or equal to your combat, about half the time you will be better off training with Sentri first (the other half being indifferent).

Likewise, training with Bradman (+2 Dex) than training with Markus will not be equal to training with Zella then training with Sentri, but it will be equal to training with Sentri then training with Zella.

I hope that I am making sense here...

Hm. Making some further analysis turn up some interesting results. For example, from the avatar's starting stats of Dex 18 and Combat 14, it takes at least 15 training points to reach 30 in both -- e.g., by training Dex to 30 then training Combat until 30 (there are other ways to achieve this optimal result). So in a way, it can be better to train Dex first then fill up by training combat... contrary to what I said earlier. Likewise, Spark's initial 22/10 can be taken to 30/30 with only 12 skill points by training his Dex to 30 then training his Combat to 30.

If you care only for Combat, the results may be wildly different, though.

I think I will make a php version of the algorithm I used to calculate that optimum and post it in the web... maybe someone will want to use it, I who knows. As if the game weren't easy enough

[weirwood]

Thanks for the input, Marzo.

Hmm, IIRC, the formula for combat gain from Dex training should be simpler, something like +C/D combat. I know for certain that once your combat value is higher than your Dex, training with Sentri will improve combat by at least +2. Which sounds like a good thing, but once your combat value reaches 30, you can no longer train with 'pure' combat teachers, who charge skill points for the combat increase.
This includes Menion, who unfortunately is also the best teacher for strength.

One thing you disregard, training Dex will always also increase combat. It shouldn't make a difference wether (for example) Dex/Com is at 24/20 or 18/14, should it? I believe only the net difference between dex and com plays a difference regarding combat training. Taking Spark, if you max Dex first, he'll be at 30/18. If you then train with Markus, he'll go to 30/24, then 30/27, then 30/29 and finally 30/30.
If you train 4 times with Markus right from the start, he'll go from 22/10 to 22/16, 22/19, 22/21 and 22/22. No difference.

[weirwood]

hmm, I don't think there's a single formula that fits all trainers. I just tested something: I had Iolo train with Jillian and Denby.

With Jillian, he went from 20/0 Int/magic to 21/10 for a cost of 2 skill points.
With Denby, he went from 20/0 to 21/11.

[keldorn]

Yes its a simple game but we want to get the most out of it for the love of it. Any luck in finding out the meaning of trainers being "extensive"?

[marzo]

The formulas are the same -- but looking at the usecode, it so happens that Jillian and Denby train Int and Magic in different orders. That mucks things up... but the formulas still work:

Int 20, Magic 0:
* Training Int first then Magic: Int increased to 21 first, then Magic increased to (21 + 0 + 1)/2 = 11
* Training Magic first then Int: Magic increased to (20 + 0 + 1)/2 = 10, then Int increased to 21

For the record: With the exception of Chad, all trainers train Dex before Combat. Likewise, with the exception of Jillian, all trainers train Int before Magic. I hadn't noticed that this happened in the usecode before or I would have mentioned it...

[marzo]

Hmm, IIRC, the formula for combat gain from Dex training should be simpler, something like +C/D combat

Like I said, I pulled the formulas straight from usecode. If I had simplified it a bit, I would have seen that the formula for increase of Combat when training Dex is just 'Combat + 1 + (Combat - 1)/Dex', rounded down, for each +1 to Dex -- basically, +1 to Combat for each +1 to Dex unless Combat is greater than Dex. Not having done that simplification (a dumb thing, I know), I proceeded to make graph analysis for several different pathways to see if there was a difference; I wrote a short algorithm which enumerates the shortest paths to maximize Dex and Combat; etc. The conclusion I reached in the end was the same one you did. But with a lot more work.

All because I basically went on analyzing instead of thinking

Which sounds like a good thing, but once your combat value reaches 30, you can no longer train with 'pure' combat teachers, who charge skill points for the combat increase.
This includes Menion, who unfortunately is also the best teacher for strength.

That last statement is actually wrong; you can still train with a trainer as long as at least one of the stats he trains is less than 30 -- you can still train your Strength with Menion if your Combat is 30. This is as true in Exult as it was in the original.

One thing you disregard, training Dex will always also increase combat.

I didn't. But in my analysis, I was too focused on the low Dex/moderate Combat portion to see the big picture -- in that region, training Dex has huge effects in Combat (e.g., Dex/Combat 2/19 goes to 3/29 costing a single skill point). I know that those cases aren't likely to happen at all, but so what?

@keldorn: I think I figured it. Here is an excerpt of Dan Simpson's trainer table, containing only the trainers you mention:

Code: Select all

  Bradman  - Yew      - Dex +2                    - 30 gold (15)
  Zaksam   - Vesper   - Combat +2, Str +1         - 40 gold (13.3)
  De Snel  - Jhelom   - Combat +2                 - 40 gold (20)
  Inforlem - Terfin   - Str +1, Dex +1, Combat +1 - 50 gold (16.6)
                        Int +2, Magic +1          - 50 gold (16.6)
  Perrin   - Yew      - Int +2, Magic +1          - 45 gold (15)

Note that -- with the exception of Perrin -- all trainers in the list give a +2 in the attribute you listed in parenthesis (BTW, shouldn't Perrin be also listed under Intelligence in our list? He even charges less that Inforlem for it...). Note also that Perrin's and Inforlem's +2 to Int will guarantee an extra +1 to Magic based on the formula I gave earlier (so... shouldn't Inforlem be also listed under Magic in your list? :-p).

These trainers are also the *only* ones that give a +2 to an attribute.

[keldorn]

Yes, I made the mistake on Inforlem earlier cos was typing from memory.:p Oh so the meaning of 'extensive' is that they increase the attribute by 2. I had thought that they give some extra bonus.

[keldorn]

Just thought of something, if extensive implies +2, then shouldnt Menion be listed as an extensive Strength trainer? The cluebook certainly didnt state so.

[marzo]

He should. Hm. Are you sure he wasn't listed?

[keldorn]

Yes I'm very sure Menion wasn't listed as extensive cos I have the cluebook. Also, the trainer list in the below link by ganesh also didnt state him as so.

http://db.gamefaqs.com/computer/doswin/ ... _vii_b.txt

He mentioned that trainers with the * gives more training for the buck, I guess he means the same as extensive, but also didnt list Menion.

(*) = Marks trainers that give you more training for your buck

[marzo]

Yes I'm very sure Menion wasn't listed as extensive cos I have the cluebook.

Hm. I was just making sure as I don't have it myself.

It is possible that Ganesh went by the cluebook; and it is also possible that Menion was overlooked in the cluebook as well -- he does give a +2 increase to Strength, and is proportionately cheaper than most other trainers. If that doesn't qualify as "extensive", I don't know what does. Especially since the only difference between trainers is the number of times they train each stat (and sometimes, the order in which they do it). And yes, I checked the usecode of each of them to make sure of this.

[keldorn]

No matter, thanks for all the checking.

[Milt]

I was looking over the formulas and I had a question. You mentioned that training combat/mana when its controlling attribute was equal or greater only increased by 1. Does that mean that if you train dexterity it will still increase combat?

I ask because from what I can tell, when combat is greater than dexterity, increasing dexterity raises combat more than it increases dex. I tested this via a simple graph of the equation. Tad confusing since it was mentioned that if combat is equal or greater than dex than its better to increase dex until it catches up, which can't happen until combat reaches 30.

[marzo]

You mentioned that training combat/mana when its controlling attribute was equal or greater only increased by 1.

Read what I said again: when training Combat/Magic, and the value given by the formula is not smaller than the controlling attribute, then Combat/Magic will be increased by 1 instead of the formula (emphasis added).

Does that mean that if you train dexterity it will still increase combat?

Yes.

Tad confusing since it was mentioned that if combat is equal or greater than dex than its better to increase dex until it catches up, which can't happen until combat reaches 30.

True. It turns out that if your goal is 30 in both, it is better to not let Combat go over Dex and train Dex to 30 and Combat to Dex in whichever order you prefer.

[TdI]

Just wondering, what is the difference between Dex and Combat anyway? I know that Dex makes you move and hit faster in combat, and I think that in the original, Dex highly influenced the chance of successfully picking locks. But what does Combat do?