Weaponsmith Archive

Ok, here I am crafting a schematic for a DX2. I have 10 Experimentation points.

I start doing my experiments 2 experimentation points at a time. I've always done this.

On this particular DX2, I got it up to the "max", and got 2.5 speed.

However, at one point while experimenting, I got a "moderate" success in there somewhere, and when I finished using up my points, I ended up with a 2.4 speed DX2, but with rather poor stats in damage and wound chance.

So, this gets me to wondering. I go ahead and attempt another DX2...this time, I use 6 of my Experimentation points at once. I get a Great Success. I go ahead and put in 2 more. Another Great. I put in the last point possible (out of 2 left). Another Great. Result? DX2 with the exact same damage and wound chance as my original schematic, BUT, a 2.4 speed, not 2.5.

This perplexes me. On every item I've made to date, it didn't matter HOW I spent my Experimentation points as long as they ended up at the same place. However, I've clearly discovered here that HOW I spent my points in this particular case MATTERED.

My reasonning is that there are more decimal digits than are displayed on the screen. The display either truncates or rounds off the ACTUAL speed value. In this case, the variance in how the exp. points were used caused the display to show different values. Maybe my first DX2 was really 2.501 speed, while my second DX2 was really 2.499 speed?

Anyone have any ideas? And yes, I've heard "On some weapons, it matters, on others it doesn't", but nobody's bothered to ever give any concrete examples so far.

its just a screwball moderate success

sometimes they improve one stat a lot while royally messing up another one

sometimes they drop 100 dmg off the gun

just stupid crafting system bugs

D'Tox, when you say you got to the max, do you mean you hit the max % of experimentation goodness possible with those materials? Like, 80% experimental damage, or whatever?

If so, that is indeed pretty weird. My only theory would have to be with rounding errors on the experimentation. you experiment some, it decreases speed by .16, but then that gets rounded to either .1 or .2, and then you experiment again, and that effect stacks... etc.

If the modified stats aren't being tracked out to seven decimal places, one strong possibility is that they're not being rounded, but truncated, which is effectively the same as rounding down every time (and much faster for a computer to do).

---

DToxWeaponsInc wrote:

My reasonning is that there are more decimal digits than are displayed on the screen. The display either truncates or rounds off the ACTUAL speed value. In this case, the variance in how the exp. points were used caused the display to show different values. Maybe my first DX2 was really 2.501 speed, while my second DX2 was really 2.499 speed?

---

Itdoes track multiple decimal places - projectile pistol barrels made with a particular material would display four decimals in speed, and there's also the thing where scopes experimented to 0/0/0 HAM will add 1/1/1 to the final product. I can't tell if it's rounding or just truncating, but it's definitely doing one of them.

Maybe the DX2 thing has something to do with the percentages on the experimentation bar. If the stats are determined by the bar (just an idea, I don't think there's any way to test this) and it's rounding percentages, there would be combinations of results and points that would yield an "extra" percentage point and thus increase stats a tiny bit.

Or maybe it's just another unexplainable oddity in the crafting system. I remember making scout blasters before I got master, and the damage would vary by 1 or 2 on different batches even though I was using the same materials and the same method.

Personally I think the answer is in making powerups. If you have experienced making powerups then you'll know the following -

• Using all 10 points at once will give significantly different results than 1 point at a time. Using 1 at a time, will generally lead to a 33%/15%/7% powerup, but all 10 at a time generally never gives the third stat (and sometimes no second)

My guess (yes it is just guess) -

• Each experimentation has a probability of affecting the 3 stats (speed, min, max). Most of the time it affects all 3, but sometimes it does not.


• Ever noticed that even though you have maxed experimentation, you still have a gem left? When you use it, it changes only one stat to bring the gun back to your expected max.

So it is a trade off. 10 experiments, which has a 10 times chance of moderate success, or 1 experiment which may not hit all the stats. Personally I go 3, 2, 2, 1, 1, 1 or 3, 2, 2, 2 ,1.

It's just screwy imo.

I have had moderate successes that were much better than my amazing successes. I dont even look at those that much anymore- i just make 5 guns up and choose the best of the lot- then i shoot for a shematic with those numbers.

From my experience with experimentation I have always had better "end" weapons with doing all of my experimenting in one shot. Ie using 6 on damage and getting a great success and then use other 4 on efficiency and get another great success. I have tried the 1 at a time and 2 at a time and 3 at a time.Even if I get all great success in the others.

I too seem to always get better results hands down using 1 point at a time. I've tried to nail down some sort of pattern or explanation but the best I can come up with is some sort of rounding. For example it seems that each great success typically increases the percentage by 7%. Sometimes it's not, but more often than not, it is 7%. What if under the covers its really 6.6%, rounded up to 7%? Taking this view, 2 stabs at 1 points each result in a 7% rounded + 7% rounded = 14% overall (assuming great successes). However both at once which would result in (6.6+6.6=13.2) which is then rounded - it's 13%.

5 points at a time would result in 33% improvement, versus 1 points 5 times in 35% improvement.

I'm not saying I have hard figures to back this up, nor am I saying this is a hard and fast rule. It just seems to more or less follow this pattern in my experience. Sometimes I will have identical percentile experimented but have different stats. Perhaps the stats also get rounded on the fly, and are not strictly speaking really based off the final percentile base? I reckon not even the devs know what's going on!

Go figure!

The difference between spending experimentation 1 point at a time vs all at once is the same as the difference between compounding interest, and applying the interest in one calculation over the whole time period.

For example: A great success increases the value by 7%.

If we have a starting value of 50, then 1 point gives 50 x 1.07 = 53.5, the next point gives 53.5 x 1.07 = 57.245, and the third would give 57.245 x 1.07 = 61.25215 if you get a great success on each.

Now, if you spent those points 3 at a time you get, 50 x (1.21) = 60.5 after a great success.

Making something 2 days ago, I could get 93% damage spending 1 at a time, and only 91% spending all at once.

Honestly, it sounds to me like the first gun you ran, doing two-at-a-time, had a hidden mod success, and the DX2 was supposed to be 2.4 all along.

I have two other Master smiths I work with, we all use the exact same materials and we all have always came out with the exact same guns (unless of course, the fluke more-experimentation-points-than-normal instance occured). I use two-at-a-time normally, and the second uses two-at-a-time and the third uses varying methods. None of us have ever, ever experienced a "method you use on some gun matters" situation and I honestly doubt such a situation exists.

Dtox did anything shift minutely? Maybe a diff type of Iron with just a LITTLE bit higher oq? or something?

Was just a thought as the stats of resources are 4 digits long, who is to say that the decimal doesnt track further than 4 digits?

I know shifting to a new type of copper that had 20 more OQ net'd me 1 extra point on min damage.

Hope it helps.

I think the crux of the matter is that all 'great' successes are not alike. Depending on the roll it may be better or worse than your last 'great' success.

The ideal scenario would be to use all of your points in one shot, and get an 'amazing' success. But that is rare (or never) because of the 'risk' in using more points. The 'one box method' is best simply because the chances of getting an 'extra good great success' is better because of the less risk.

You got a little extra speed on that one because you used 6 points and got an 'extra good great success'. Just my2 cents, Icould be wrong.

I made a few batches of guns yesterday using 3 at a time instead of my usual 2. Number of points required to max the bar and final stats were both unchanged.