T2A skill advancement modified by server population?

[Excalibur]

I was wondering, is skill advancement modified by server population? I.E The more people on, using the skill, the slower the advancement? I know this definately applied to blacksmithing, but it doesn't seem to be documented for the rest of the skills, so I have no idea if that's how it worked for the rest. Food for thought anyway, if it's working that way, great.

[Orsi]

I for one would like to know this also. GM Blacksmiths, Bowyers, Tailors and other craftsmen were few and far between, and definitely not something which someone had an alt for each.

I believe it was like that, I think I skimmed over something pertaining to this in the many sites I've read of the era. It probably is the case on the real servers now too.

[Excalibur]

Yeah, I know this was a major source of frustration for blacksmiths at the time (I imagine it was the same for most crafting skills), and even prior to T2A. I found Slyfox's older guide that was on Stratics over at the web archive, and also cross posted on the UBB site at the time. You can find them here and here.

The main point being:

The time, shard and server you play on has an effect on your skill gain. Skill gain is best when there are as little people on the same server using Blacksmithing skill. Thus European times are much better for gaining skill than prime US-times. If you however do play prime US-times on a US server you will need approximately 5-10% more ingots.

I remember this too, because it caused quite a stir after he posted it, I had hundreds of emails pouring in complaining about the general randonmness in smithing from other players. Despite repeated attempts, neither DD or Carly Taylor were forthcoming about why that was or how it even remotely worked. It was the one thing, probably the only one, that I could never nail them done about regarding smithing.

[Derrick]

The skill gain rate on UOSA is not modified in any way by other players. I do understand(mostly) the system that OSI used in which there was essentially a skill point pool and each player online would draw from that pool when attempting to gain. This is acutally very evident in the T2A demo, where skill gain is incredibly fast, with only one player online.

This is something that we've discused implementing, and other emulators such as POL did have this type of system implemented.

[Excalibur]

The skill gain rate on UOSA is not modified in any way by other players. I do understand(mostly) the system that OSI used in which there was essentially a skill point pool and each player online would draw from that pool when attempting to gain. This is acutally very evident in the T2A demo, where skill gain is incredibly fast, with only one player online.

This is something that we've discused implementing, and other emulators such as POL did have this type of system implemented.

That would definately be cool if it could done. I think it would pay off in the end, when the server population gets much higher, especially for the crafting skills. It would encourage less mules and make a genuine blacksmith more valuable then they currently are. It'd be fantastic to see the Britain forge alive with activity again, as it used to be.

Thanks for the quick reply Derrick, I know your busy.

[Faust]

I am a bit curious what the values of the numbers in the pool would be if this was ever implemented. OSI servers clearly had thousands of people, which would require a much larger number compared to here. This type of project would take a lot of trial and error testing to get it down just right.

[Kaivan]

The comments at the bottom of this page by Designer Dragon should help explain the general concept of the server population effecting skill advancement.

[Faust]

I have read this before but something struck me while reading it again this time around...

The server tracks the # of tests called on each skill since the start of the shard.

It tracks them, however, as a PERCENTAGE of total skill calls.

Also, there's a cap; if some skills look like they are going to dominate the percentages and squeeze out the others, they're not allowed to account for more than, uh, say 10% of the calls each.

The advancement rates are then based on these percentages. The hypothetical average player who plays "typically" would see all their skills advance at exactly the same rate in a given span of play time. Of course, nobody is the average player and nobody plays "typically." The result is that for any given player, some skills seem harder and others easier.

Take a note on this section of the article and specifically the underlined section... Now if you look at an outpute file that OSI generated for stratics a long time ago in '98 for skill specifics it gets a bit more interesting...

Here are the first two skills int his output file below...

version 1
StatCurve 1
SkillCurve 250
################################
Skill: Alchemy
Strength: 0
Dexterity: 50
Intelligence: 50
StrReq 0
DexReq 0
IntReq 0
AdvRate 8000 2000 100
StatAdvRate 0
SkillStat: 90
CanUse: FALSE
SkillWeight: 1
SkillScript: alchemy
Version: 15
################################
Skill: Anatomy
Strength: 15
Dexterity: 15
Intelligence: 70
StrReq 0
DexReq 0
IntReq 0
AdvRate 2000 500 25
StatAdvRate 0
SkillStat: 100
CanUse: TRUE
SkillWeight: 1
SkillScript: anatomy
Version: 13
################################

Again, take special notice to the lines underlined...

If you take a look at the skill specifics it actually lists the advancement rates. This can possibly refer to the number of times a skill can advance in 3 particular areas. That starting at 33.3, 66.6, and the rest. I simply divided 100 by 3 to get these values. This was a bit interesting because skills always would seem a bit more difficut in the 30's and closer towards the 70's...

This possibly could mean that less than 33.3 the pool was 8000 for alchemy and 2000 for anatomy, 33.3-66.6 it was lesser resulting in 2000 for alchemy and 500 for anatomy, and 100 for alchy and 25 for anatomy for the rest. The output file lists every single skill and this could possibly give us one thing we have been missing for a long time now. That being era accuracy for skills.

[Orsi]

I am a bit curious what the values of the numbers in the pool would be if this was ever implemented. OSI servers clearly had thousands of people, which would require a much larger number compared to here. This type of project would take a lot of trial and error testing to get it down just right.

I don't think it would necessarily matter about the size of the population. I'd imagine it was proportional to how many people were using a certain skill at one time to population size (something like that). More people at peak times = more people using blacksmithing = less skill increase. Less people at off times = less using blacksmithing = more skill increase.

It would probably raise and drop difficulty by relative server sizes, I wouldn't think they hard coded the absolute numbers for each server size.

Edit: Just read your above post. Funny thing is, I was looking at that skill list yesterday and seems like a coincidence seeing it today.

[Excalibur]

Those are some great finds guys, nice work. I dropped Raph (Designer Dragon) a note asking if he might have some more info that would be useful regarding this issue:

From: "Raph Koster" <[email protected]>
To: <[email protected]>
Subject: RE: Message from BlueMail
Date: Monday, April 27, 2009 11:02 AM

Hey there, good to hear from you!

I don't have any of those old emails, alas, and I also don't have any code
or formulas. and my memory for 15 year old numbers is quite iffy. So I
don't know that I can help, honestly!

So unfortunately, that is a dead end. I'm pretty sure it wasn't based simply on the people using the skill at the time the skill check was made, but additonally further checked against all players on that shard, that had any skill in said skill and then pooled for a final skill check. The more people who had points in a skill, online or offline, using or not currently using, made it dramatically different.

I put together a study group after Sly Fox published his guide because we had some suspicion about this, one group started from 0 blacksmithing skill on Atlantic, the other started on one of the new CST servers (I can't remeber which, for the life of me) and both groups were instructed to advance to GM during off-peak hours only. The results were never fully completed because OSI threw in yet another of their stealth skill difficulty changes and screwed up the whole study, but with what we did get,we found about a 10-17% difference between Atlantic and the new server.

At any rate, it's definately going to be impossible to reproduce it precisely as OSI had, like you said Faust, but yeah, trial and error would be good enough if this doable.