Another derailed thread 
The info on that site isn’t very good, look me up on there. The problem with the info is, it depends on players uploading their combat logs to them. The data is really rough and doesn’t get updated outside of player uploading logs.
I’m aware of that and I wouldn’t use it to make a judgment on a players ability unless the sample size is at least 100-150 uploaded but thats all you got to look up players like gansterism who got their profiles hidden, some data is better than no data. Now you think 100 games might not mean much but it do. Lets say your using a hud in poker the rule of thumb is about 300 hands but you can make really accurate judgment calls with as little as 50 to 100 so I apply the same rules to crossout info.
I mentioned some of this when the test server was live. I didn’t think it was that bad though. I had said you have to be careful when bringing the guns around especially coming out of turns as I was missing more shots and using more ammo then I expected too use.
What they didn’t do to shotguns that they did to the other guns while switching over is applying compensating changes to accuracy. It’s in there patch notes:
“For all other weapons, except for shotguns (because they would be minimally affected by the change due to the low range, which we do not plan to increase), we have added compensating changes related to maximum range and accuracy, which should improve the game experience while increasing the requirements for the player’s skills.”
So it’s likely in these fields that the problem sits:
Min. spread, Max. spread, Rate of spread, Spread…
(These were harvested off of the changes to other guns and there are some other spread fields too.)
It doesn’t seem like it’s a bug more so they just kind of ignored working on them do to the assumption that their low range usage wouldn’t be as affected as much.
I normally would agree but when I was looking at myself on the site I was thinking bad data could also be very misleading. If I remember correctly that site was hoping players would upload their own logs to get statistics out of them in return. Everything else they collect with it was kind of incidental info.
1 Like
I would make an argument it is even more accurate in a way than the official in game stats for one solid good reason in that it is very current. For example my average kills per battle on my main account is 3.19 but my much newer alt account is 5.4. Players improve over time and I don’t think that site has been around for very long but I am not sure.
You only got a 8 battle sample and even that can give you some idea of a players ability for about 8 out of 10 of players so you can make a loosely accurate judgment call of a players ability 85% of the time with even that small a sample size while the rest would be way of the mark and no where near the ball park. With 100 games the odds of being way way off the mark(lets say 10% for example) are next to zero and should be within just a couple percent (5% max) of a persons true stats.
The sites been there for a number of years though I think it was down for a bit.
It’s not going to give much, unless a lot more samples were taken. You would want like that 100 samples you mention for each build/weapon and mode a person plays in for it to be fairly accurate. That’s unlikely to happen unless they work out some sort of deal with the devs like crossoutdb did.
Right now they only have really limited and random dated data sets to work with. That’s not really good for accuracy.
True to an extent but you can stil getl a fairly accurate picture of a persons skills with a large weapon sample size, you don’t need every weapon and if they can’t do good with even their specialty weapon than they most likely suck(mine is porcs). Players got better or worse win rates with different weapons thru trial and error and gravitate towards the weapons they do best with after lots of experimentation and some players just do good with everything because they are very skilled
Userlol for example got varying win rates with different weapons but there is one thing in common and that he does fairly good with all of them. His pvp win rate is also lower than cw win rate most likely due to playing solo from time to time in missions. I beat him 3 out of 4 games just the other day but I was using porcs which is my specialty with a solid 65% win rate and thats just playing solo to make the best of it. I am the top porc player for missions this season which is prob where that data came from cause i say it is only the serious hardcore players such as userlol uploading their logs to crossout info but it is still somewhat useable info. Like I said if players can’t do good or at least break 45%-50% winrate with their specialty weapon even as a solo player like myself then they prob suck with the rest of it as well.
But yea 100, 200, 300 games is still somewhat accurate and you can still paint an image of a players ability even with that small sample even if they don’t have a wide weapon variety under their belt because most players play to win and they ain’t gonna play 100 games with a weapon that they suck too too bad with. Also most of this data comes from hardcore CW players so if you get a battle logged it is prob in a serious match at high ps where most players are playing to win.
Edit: No i don’t go rubbing peoples noses in it cause i could care less but when theres players in game(most of them with hidden profiles) trash talking me or fellow teammates then crossout-info gives me an out to find out if they are full of xxxx or not so I can put them right in their place in the game chat and they always usually go silent as soon as I shine a spotlight thru their bs.
YES!
i was experiencing this heavily with fafnirs in pvp yesterday. i was running two of them but id constantly be missing my shots and even when they landed it felt like they were doing no damage at all. i was being beaten by bots of all things. FRICKEN BOTS!! now its rare that i lose from full health to a bot but omg the bots were so irritating. id get right up on them and shoot but sometimes nothing would happen or my fafnirs would do little to no damage. mind you, this is with two fafnirs on a jackie so they shouldve been doing more then nothing. id be aiming right on the weapon and fire but it seemed that my shots would miss at point blank range. i even had the full 40% boost on them and still it was bad.
read above.
yes this was an issue with fafnirs as well. i tried using them for a few rounds but good lord man they seem to not be accurate or hit anything.
yes!! i was playing them yesterday and even with the boost from the jackie they SUCKED really badly. good lord were they terrible. i was wondering how the hell i was getting my ass handed to me so easily by the bots when im running two fricken fafnirs. i was going to make a post yesterday about it but im glad this is confirmed that im not just thinking wrong here.
because its fun to have a variety of weapons and whatnot to choose from.
1 Like
I am glad I do not own any of these shotguns.
I think you figured it out.
Seems like they underestimated how the old values would carry over when converting to projectiles.
Although it almost feels like the inertia mechanic of Yongwang and Porcupines has been applied to them. I’m not noticing the same thing with MGs.
Playing shotguns on Jannabi used to be fun, but now it’s like turning and drifting throws your accuracy way off the crosshairs.
The oldest games you can possibly find dates back to February 2023 and I was surprised it even dated that far back so I don’t know if it is due to the down time you mentioned or they just hold unto logs within the past 2 years. A trend that I noticed is that the amount of daily uploads dwindled the further back you went so it seems that quite a lot more players are uploading their logs these days so anything on the site is most likely recent.
its better to have no data than incomplete data.
what you see completely depends on who decides to upload their logs on there
You can kind of compare it to using an hud in poker. All the data is incomplete and will never be truly accurate but it is accurate enough to get a rough picture of a players playstyle, ability and what not so yes incomplete data is better than no data. Hell you can get even a rough accurate guess of a players pre flop raise percentage stats with as little as 7 to 8 hands which is as low as a single complete rotaton of the button. I’m pretty good with odds if there is one area of math I’m kind of good at it is combinatorics.
And yea it was accurate enough to know that you only play mammoths before i even checked your builds on the exhibition so it is good for something.
Yes i say the odds that players uploading there are more often than not good players so you can make a slight adjustment by raising a few things a couple points higher than what they are which is called weighting, election polls does it for example to adjust for unseen variables like the notorious elusive trump voter lol.
There could be any number of reasons. I know at one point the site that use to do similar the domain was up for sale. So it could be 2 different similar projects or the same one. If it’s the same one reopening they might just not have kept the old data as the data/time gap could create problems.
The only thing I ever used the old one for was looking at weapon rankings. That’s probably more accurate than the individual player stats.
Depends on how you are looking at it much like the video you posted was showing you can have 3 builds all using the same weapon and get drastically different results with it. Remember JB didn’t like the second build. That’s going to reflect faster in stats with a small sample size then it would with a large sample size.
If your using the site for yourself to track your own progress it will work great. However it’s going to be inconsistent data for people that aren’t providing their own logs. Those inconsistencies can swing wildly too as I mentioned before the smaller the sample size the larger the rate of change in the stats can be seen. So you might see one player using a weapon where the data says they are bad at it but they are actually good just from the inconsistency of the data.
Someone should try doing a spread minimizing build with the fafnirs to see if it helps. As not having them be range adjusted could also be playing a part in it. Most of the MGs got additional range in the change over. So it could be an issue of both as well.
1 Like
the incomplete data gives you the wrong perspective so you would be better off with no data
thats the only thing its good for. if i played any other builds that werent uploaded by anyone, youll never see them
i would call that a bit more than “slight adjustment”. only total sweats would go as far as to upload logs to this stat website
It is definitely good enough for ruling out the extremes. No it is not accurate to the exact percentage point but that ain’t necessary. The extremes would be between around 30% to 70% because no one is losing 30% of games in pvp even if they were afk because you got other players on the team as well. 30% would be the bottom of the bell cure and i say 80% would be the top so at least 95% of players would fall between 40% to 60% win percentages and the truly skilled players would have a win rate of at least 60% minimum. A 50 game sample size is enough to determine if said player is outside that bell curve.
With regards to weapons the majority of the games will be logged with a players primary weapon that they spent the most time homing their so called skills on and if i see a player is just mediocre at that then they ain’t skilled they are just that mediocre at best or absolutely suck at worst. Also most players ain’t skilled but just average, like if you ain’t got a win percentage of at least 60% than you got no right to be saying what takes skill or don’t take skill.