Pistoleer Archive

I've been trying to analyze the accuracy of pistols given the latest reduction in accuracy for pistols --trying to understand which pistols to pick based on accuracy over range. I have some initial conclusions here, but need a lot more data to back them up.

Jaegen88 -- we may want to consider putting the hosed range indicators on our issues list. It's almost impossible to verify results when the accuracy indicator seems to cap at -50, but the weapons stats go as low as -90.

Since direct observation is out of the question (due to the inaccuracy of the mod indicator), I've taken my most frequently used pistols and graphed the Min/Max/Ideal modifiers then interpolated for the ranges between these points. I then tried a number of different curvefits to try to characterize the nature of the accuracy formula. Results so far have been disappointing.

I expected to find that a relatively simple 2nd order polynomial would describe the accuracy calculation curve. This does not seem to be the case. I also tried a Logarithmic fit, with even worse correspondence to known data points. My current model uses two slopes (Min to Ideal and Ideal to Max) to characterize the accuracy performance. Intuitively I don't believe a developer would have taken this approach, but I have no data to support another conclusion

Taking into account those limitations here are my conclusions:

o

Best chances are best chances. RNG or no RNG. You use the weapon with the best damage range (and a good speed) that you can find, and sure the RNG could give you low damage on any given hit, but best is best and will prove to be so over time.

Anyway - Actually I DO think that it is probably as simple as two unique extrapolation lines. That is very easy to program (not that a second order polynomial is hard to program... one conditional instead of a touch more calculation - simple either way). But most of the other equations we've been able to figure out so far in the SWG game mechanics have been very simple, so that would fit the bill.

I must say that my Advanced DX2 brings down animals and imps alot faster than my FWG. Plus with accuracy boneses in alot of pistoleer/BH skills...i dont think it really matters. My weaker DX2 hits for 136 top end....my FWG hits for 203. Speeds are the same, but i bring stuff down alot faster with the DX2. Maybe its tha armor peirceing....who knows. All i know is that i love it!!!

---

Paracelsius wrote:

The number DOES change for movement and for posture. I think there may also be a different accuracy rate for walking v. running.

---

Definitely. I believe it takes into account your movement (slower raises acc), the mobs movement (slower raises acc), your posture (lower raises acc), the mobs posture (higher raises acc) and range (closer raises acc).

Come to think of it, maybe it does include weapon accuracy mods. I'll try to check tonight....

The likely reason your DX2 is killing much faster is that more mobs are vunerable to acid based attacks then heat attacks. If they mob is vounerable to that attack then you bypass all of its armor. I am not sure but I think the damage numbers you see are pre armor reductions. Can anyone verify this?

I was wrong. The accuracy number definitely takes into account your weapons range modifiers. I'm still not sure it takes into account your skill, though...

I'm 80% sure that your chance-to-hit modifier does NOT take into account your players skill into the equasion, as using a Scout Blaster now the numbers seem to be very similar at the different ranges than when I first was able to, otherwise 50+ bonus to accuracy would be something I would notice for sure.

Your accuracy is not taken into the +/- meter above the mob/player name. I tested this with a friend that had a lower accuracy in the weapon. Basically went 15m away, and even though my accuracy was much better than his, in his hands the same weapon showed the same number above my head.

Accuracy does improve the to-hit chance though. I am sure I miss a lot less with marxmanship 3.

I've been trying to analyze the accuracy of pistols given the latest reduction in accuracy for pistols --trying to understand which pistols to pick based on accuracy over range. I have some initial conclusions here, but need a lot more data to back them up.

Jaegen88 -- we may want to consider putting the hosed range indicators on our issues list. It's almost impossible to verify results when the accuracy indicator seems to cap at -50, but the weapons stats go as low as -90.

Since direct observation is out of the question (due to the inaccuracy of the mod indicator), I've taken my most frequently used pistols and graphed the Min/Max/Ideal modifiers then interpolated for the ranges between these points. I then tried a number of different curvefits to try to characterize the nature of the accuracy formula. Results so far have been disappointing.

I expected to find that a relatively simple 2nd order polynomial would describe the accuracy calculation curve. This does not seem to be the case. I also tried a Logarithmic fit, with even worse correspondence to known data points. My current model uses two slopes (Min to Ideal and Ideal to Max) to characterize the accuracy performance. Intuitively I don't believe a developer would have taken this approach, but I have no data to support another conclusion.

My sample was a DL44, 2 DX2's, a Tangle, 2 Scouts, an SR, a DH17 and an FWG5.

Taking into account those limitations here are my conclusions:

o The FWG5 Pistol is far and away the best pistol from an accuracy standpoint for all ranges outside of melee range (say 8m)

o The DX2 is a good point blank choice, but very rapidly decreases. Overall the DX2 is pathetically inaccurate.

o The Tangle and SR Combat seem to share a single accuracy profile. That profile is an almost linear decrease from 0 at PB range to their minimum at 48 range. An interesting question is -- are there any other 48m Max Range pistols and do they ALSO share this same slope?

o All the energy based pistols intercept the X-axis between 21 and 22 meters range. So blasters seem to be "zeroed" at 22 meters.

o For those that like to fight from just outside melee range (Creature Handlers?) a DL44 is a good choice, but over longer ranges lose out to the FWG5.

I plan to keep looking into this subject as additional information becomes available and I have access to more models of pistols. Feel free to let me know if you see any problems with this analysis or if you think of other lines of investigation.

Regards,

Oirff (Bria), Expert Gunfighter (0,1,4,2)

Yeah, you pretty much lost me when you said '2nd order polynomial'.

But if one of the ideas you threw out was the the accuracy curve was linear from PB to ideal range, has a kink, then is linear out to max range... I would find that very feasible. Those equations are fast to do, and easy for a game designer to balance. (Except for SOE designers of course </end zing>

Hmm, I thought the number by the mobs head was only taking into account posture, movement and range, and did not take into account your weapon's accuracy and your accuracy skills. As such it maxes out at 50 because they want the maximum possible penalty (regardless of weapon) from posture, movement and range to be -50.

Does the number by the mob's name really change when you switch pistols while maintaining the same range?

I'm also with EasyMcRhinopants that I figured the accuracy graph was linear between min and ideal and linear between ideal and max.

I appreciate all the work you are doing, but isn't the majority of combat based on randomly generated numbers? If so, I would think this is why you can't find a curve to fit your data. I would think that if you did 10, 10 shoot samples with 1 pistol you would get different data for each 10 shot sample.

I think you can draw some conclusions with hundreds of data points, but depending on your sample size I would be cautious in trying to place a lable of more accurate or less accurate on any one pistol.

Interesting questions SmugglerFlux. I may have a flawed assumption. I'll check it tonight and let you know.

The number DOES change for movement and for posture. I think there may also be a different accuracy rate for walking v. running.

If it doesn't take into account weapons or skill accuracy as you suggest, it makes it of limited value...

Oirff