T key accuracy report varies

[Case23]

When I use the T hotkey to target a unit, a custom test weapon reports accuracy percentages of 15% - 18%.

In pursuit of better understanding how the accuracy stat affected likelihood to hit, I created 5 custom versions of the M91/30 rifle in the Russian OOB with accuracy stats of 0, 5, 10, 15, and 20 and then units and formations armed with it and it alone.

I set up a human on human battle between Soviet and German forces, bought the appropriate five Russian units, left the Germans with only their Kommandeur, and maneuvered the units to within 50m of each other.

I used ScenHack to set all experience values to 60 for the Russian units.

When I use the T hotkey to get the accuracy % for the 0-accuracy unit, it is variously reported as 15% to 18%.

The other units (with weapons of accuracy 5, 10, 15, and 20) consistently report an unchanging value for accuracy %.

A special case for zero accuracy weapons? Something else?

Saved game files enclosed in case they're useful. Relevant unit is B0. Press T repeatedly to duplicate behavior.

Thanks in advance.

- Case

[Case23]

Andy, Don, any thoughts on this?

- Case

[iCaMpWiThAWP]

I don't think this is an issue, there is probably a random value used in the accuracy of the shot, to prevent sure hits all the time.

[Case23]

Agree it's probably not an issue. But the behavior isn't uniform. Most of the time the accuracy indication doesn't fluctuate like this (though it may change if you switch targets and lose the accuracy gain from repeat shots and the like).

This is an outgrowth of attempting to understand in more detail how the accuracy stat affects chance to hit. LDO, higher is better, but I'm interested in being able to make informed tradeoffs between, for example, accuracy and kill, in which case it matters how much more likely to hit a higher accuracy weapon is. I'm using the indicated to hit % rather than having to run statistically significant tests (which would basically be prohibitive). But if the to hit indication fluctuates, I have to consider how reliable a metric it is.

[Yes, I may like to analyze things more than the average bear. ]

- Case

[Mobhack]

The T key is really only useful to select targets with. It gives a rough idea of the percentage to-hit chance for the slot 0 weapon - which may well be a wild gunners overestimate it seems, sometimes.

The only accurate to-hit percentage is the one shown on the top info line when the weapon is actually fired and all the "dice" have been rolled. That may be rather different from the "T" key guestimate.

If it pings and pongs about then you probably are on some sort of "edge" condition, is my guess.

Andy

[Case23]

That fits. It had seemed that counted hits weren't as many as probabilistic hits but I was chalking it up to observer disappointment bias. That also explains one observation where hit probability went down with a range reduction in a test scenario.

I know the exact formulas are confidential as a condition of code release from SSI, but can you give any hints as to the relation between accuracy and hit probability for direct fire weapons?

I recall seeing a post in which either you or Don said something to the effect that accuracy had no effect on indirect fire weapons, but I think I saw another saying that one of the patches altered that so that higher accuracy tended to cluster shots within a tighter range. Is that indeed the case? Does higher accuracy also tend to reduce drift of the aim point in the bombard screen?

I did have a moment of trouble figuring out how to account for an accuracy of 0 for Katyushas in my model of indirect weapon lethality. Eventually, I decided to just ignore it.

[Mobhack]

Originally Posted by Case23

That fits. It had seemed that counted hits weren't as many as probabilistic hits but I was chalking it up to observer disappointment bias. That also explains one observation where hit probability went down with a range reduction in a test scenario.

I know the exact formulas are confidential as a condition of code release from SSI, but can you give any hints as to the relation between accuracy and hit probability for direct fire weapons?

other than what is written in the mobhack help etc - no

if you want to build up some statistics and "back fit" the results to some formula, then use a firing range scenario and fire a statistically significant number of shots in a guaranteed identical manner. Observe the reported to-hit chance as reported in the info line.

I recall seeing a post in which either you or Don said something to the effect that accuracy had no effect on indirect fire weapons, but I think I saw another saying that one of the patches altered that so that higher accuracy tended to cluster shots within a tighter range. Is that indeed the case? Does higher accuracy also tend to reduce drift of the aim point in the bombard screen?

accuracy is used in the direct fire code, and nowhere in the indirect bombardment code.

I did have a moment of trouble figuring out how to account for an accuracy of 0 for Katyushas in my model of indirect weapon lethality. Eventually, I decided to just ignore it.

cheers
Andy

[Case23]

Thanks very much.

- Case

[DRG]

The to hit chances are influenced, like a lot of data in the game , on variables. Those variables are what keeps the game interesting. Much the same affect can be observed when running the AP calculator and the AP calculator is showing the average results of 1000 simulated shots each time it is pressed and each time it is pressed you will see a change somewhere

Don

[runequester]

Maybe a related question, but for units with 2 identical weapons (typically 2 LMG slots) sometimes the accuracy displayed when you fire will be a percent or two different between the two weapons.

Is this due to the slot the weapon occupies, random factor or something else?

Thanks!