Jump to content
Sign in to follow this  
You need to play a total of 20 battles to post in this section.
Varathius

Need W/L statistics to proof that individual player action affects it

72 comments in this topic

Recommended Posts

Beta Testers
182 posts

Can someone please post me statistical tests obtained where individual actions of players do in fact influence the whole W/L statistics in terms of how good or bad a player really is based on W/L statistics? I keep seeing people at game start accusing (especially accusing T10 players) that they suck etc based on their W/L stats, so I was wondering to see the sample results out there of people that are good with statistics, so I can follow up on this? The statistics need to take into account: AFKers (especially DD AFK at start), Game crash bugs, poor player habits (going into canal at map starts/rushing into cap at game start with wrong cruisers and BBs), CVs that are AFK, etc, etc, etc.... 

Again, I do not need words and speculations, I need maths and statistics. If anyone can provide it, thanks in advance, if not, I guess I'll start have to defend a lot of T10 players when I see them in the near future.

Edited by Varathius

Share this post


Link to post
Share on other sites
1,528
Members
4,274 posts
4,649 battles

Those aspects even out in a large number of games, as player skill ends up being the most consistent thing. 

Share this post


Link to post
Share on other sites
777
[BARF]
Beta Testers
5,816 posts
5,484 battles

Can someone please post me statistical tests obtained where individual actions of players do in fact influence the whole W/L statistics in terms of how good or bad a player really is based on W/L statistics? I keep seeing people at game start accusing (especially accusing T10 players) that they suck etc based on their W/L stats, so I was wondering to see the sample results out there of people that are good with statistics, so I can follow up on this? The statistics need to take into account: AFKers (especially DD AFK at start), Game crash bugs, poor player habits (going into canal at map starts/rushing into cap at game start with wrong cruisers and BBs), CVs that are AFK, etc, etc, etc.... 

Again, I do not need words and speculations, I need maths and statistics. If anyone can provide it, thanks in advance, if not, I guess I'll start have to defend a lot of T10 players when I see them in the near future.

 

that is a very hard ask. In my experience in arena titles, whether one person can cause a win or loss is down to circumstance most of the time
  • Cool 1

Share this post


Link to post
Share on other sites
3,513
Members
16,315 posts
12,285 battles

I've improved my winrate by playing ships that I know I can play well recently, though I am nearly 6000 matches in so progress goes by months not days...

Edited by slak__

Share this post


Link to post
Share on other sites
Beta Testers
153 posts
7,170 battles

First of all; Who cares? Ignore those people because they are the idiots trying to tell everyone how to sail their ships every map anyway I bet lol. Second, if you want some kinda of validation you can just go check out the people with the highest win rates and look at their stats yourself. You probably won't find any potatoes with a high win rate with thousands of games played. I am sure there will be an exception or two but I haven't found one yet.

Share this post


Link to post
Share on other sites
1,849
[AXANR]
Members
3,650 posts
23,502 battles

Given a large enough sample size, a player's skill can be determined by their win rate. An amazing player can lose ten in a row, a terrible one can win ten in a row, because 12-person teams. But an individual makes a big enough difference that over hundreds or thousands of games, their win rate will mirror their skill. 

 

This does not need proof. This is basic statistics and probability. 

Share this post


Link to post
Share on other sites
Members
2,046 posts
2,264 battles

If by "affect" you mean very slightly influence- the proof is every stat in the game.

 

If by "affect" you mean heavily control or even primarily affect- there is none, because it does not.

Share this post


Link to post
Share on other sites
Members
1,699 posts
9,033 battles

 

:hiding:Got a lot of battles already I do play for fun..i dont give a Potato about Stats etc or even pay attention about improving or teamwork ..UNTIL i tell myself i think its time to improve..Improving is fun...till now im still trying to improve.

Share this post


Link to post
Share on other sites
Members
1,776 posts
6,776 battles

You cannot tell if someone is a bad or a good player by just looking at winrates. You need to look further and look at how many battles they have played in a certain ship together with their winrate.
​Having a guy in your team who is very experienced with his ship will help for sure, which automatically means he will win more battles over time. Can Messi win all games? No he can't, but the chance of a win is often bigger then without him.

​Someone with 5,000 games in an Erie, 60% winrate, who free exp'd to a Montana isn't a good player.................. Or he is that guy...

​Hopefully that explains :)

 

 

 

 

 

 

 

 

 

 

Edited by joris92

Share this post


Link to post
Share on other sites
244
[2GOOD]
Members
943 posts
23,091 battles

Given a large enough sample size, a player's skill can be determined by their win rate

 

Unless they mainly play in divisions, in which case their win rate is anything from 10% to 25% higher than it would be when playing solo.

Share this post


Link to post
Share on other sites
3,513
Members
16,315 posts
12,285 battles

 

Unless they mainly play in divisions, in which case their win rate is anything from 10% to 25% higher than it would be when playing solo.

 

i give people grief for Div'ing but 25% is insane, please tell me where you came up w/25%?

Share this post


Link to post
Share on other sites
Beta Testers
282 posts
4,998 battles

 

i give people grief for Div'ing but 25% is insane, please tell me where you came up w/25%?

 

My WR % in divison is actually lower for 2-ship and the same as solo for 3-ship for a long time because the people I like to play with (my friends) aren't very good. I play 4 funsies tho, so I don't care. 

Share this post


Link to post
Share on other sites
1,363
[HYD]
Members
7,105 posts
5,289 battles

I know for a fact that i can affect my own winrate in low tier ships, because i am significantly better than most players at those tiers. So even with a bad team, i can often carry them to victory. My Clemson has something like 60% for winrate, much higher than my total average of 53%. (sample size is about 250 battles.)

Once above tier 5, i am of average skill, so my performance doesn't really do much. I neither drag down my team, or carry it. 

Share this post


Link to post
Share on other sites
1,386
[SYN]
Members
3,775 posts
25,422 battles

It's easy to say Div players are being carried but for the few I have looked at their solo stats aren't that dissimilar to their div results. Usually within 5%.

 

25% I think is a ridiculous figure.

Share this post


Link to post
Share on other sites
Beta Testers
2,196 posts

Stats are only a book cover in my opinion. If their stat block is blue and purple, they are probably pretty good and can influence the game in a positive manner on a regular basis. If their stat block looks more like a German flag, they probably do not have a good grasp of the game mechanics or tactical awareness. If their stat block looks like a brick, its probably a screenshot of the contents of their skull. Nobody is completely undeserving of their book cover but if you reach for a copy of "Star Wars", beware the cover with an old grandmother knitting a sweater. These are just my observations, you can check Warships Today to look at the top players and their stats. Me? I'm "better than average". You can watch this old video someone made of one of my replays and check my stats for yourself, Warships Today saved an old snapshot. 

 

 

 

 

Share this post


Link to post
Share on other sites
520
[-FBS-]
Members
2,646 posts
4,290 battles

It's easy to say Div players are being carried but for the few I have looked at their solo stats aren't that dissimilar to their div results. Usually within 5%.

 

25% I think is a ridiculous figure.

 

Maybe most Division players aren't the hardcore type? 

Share this post


Link to post
Share on other sites
725
[NEUTR]
Members
2,207 posts
11,692 battles

Yes player action affects outcome.

 

You can think of it intuitively. "the only constant is the player", is the phrase often quoted.

 

The more players in a match, the more varied the map and game modes, it only increases time to converge to your true "skill rating". But as long as you play enough matches, it is only a matter of time.

 

Your rating, however is not a constant value, it is a range. You go up and down within that range approaching your true rating but never actually obtaining it.

 

The rating forumula used by warships today is simply comparing your stats with the average, higher stats you have, higher rating you are. Rating is just a number indicating your percentile.

 

If the rating distribution is normal, and loks like it is, a rating of 1000 is actually very good, you are slightly above average putting almost 60% of the players below you. 

 

Warships.today basically collects all the stats from the player pool. Find the distribution of the stats, then check what stat you are and what your stats below in that distribution. They fudge the number a bit with different attributes, but overall it's a reliable measure. 


Given the context, if a player consistently out performs the average, then he is by definition having influence on the outcome of his games.

Edited by NeutralState

Share this post


Link to post
Share on other sites
Beta Testers
182 posts

Some interesting answers...I agree with pretty much what you all had to say. I just wish that people would generally speaking, be a bit more careful with immediately judging other players over W/L. It is what I thought, there are too many inconsistencies in the game where each individual can be judged if their skills are good or bad, based on W/L alone. For instance, for DDs imho it would be ideal to show how much ships spotted, cap point capped and defended, how many other ships that were spotted got immediate DPS on them, etc, but such statistics, nowhere to be seen. A DD that does great scouting and capping does too many times lose because of: AFK players, conga lines, people using map wrong (rushing through canals for instance), etc...

I personally think a lot of the hate would also go away if W/R would not be such a primary concern. In WOT it is way bad and out of control I think. But then again, after a certain experience or many matches, I am sure that high experience players will contribute more to a game, but only over time. Still interesting to see that the guys that accuse others of being bad in terms of W/L observations, they themselves are not those experienced individuals. 

Share this post


Link to post
Share on other sites
Alpha Tester
1,021 posts

This is why I block as much as possible and think what helped kill W.o.T for most is XVM. YOUR stats should be open to WG and yourself if that's what your choice is.  There is ZERO reason outside of someones ego that anyone else needs to even remotely know someone elses stats.  Yes I am one of those folks who has as much blocked as possible and think that there is to much information going out. Someone being harassed because of their lvl of skill or similar reasons should be something should WANT to avoid.  This game can be very frustrating without having to deal with folks like that or putting folks in a position for them to be embarrassed by others trying to shame them about their stats. 

Share this post


Link to post
Share on other sites
1,849
[AXANR]
Members
3,650 posts
23,502 battles

 

Unless they mainly play in divisions, in which case their win rate is anything from 10% to 25% higher than it would be when playing solo.

 

Most of the people I div with are better than I am, and my win rate for divisions is within a couple points of my solo rate. I think the server average for division win rate is like 1.5% higher than solo. 10 to 25 is ridiculous. 

Share this post


Link to post
Share on other sites
1,528
Members
4,274 posts
4,649 battles

Some interesting answers...I agree with pretty much what you all had to say. I just wish that people would generally speaking, be a bit more careful with immediately judging other players over W/L. It is what I thought, there are too many inconsistencies in the game where each individual can be judged if their skills are good or bad, based on W/L alone. For instance, for DDs imho it would be ideal to show how much ships spotted, cap point capped and defended, how many other ships that were spotted got immediate DPS on them, etc, but such statistics, nowhere to be seen. A DD that does great scouting and capping does too many times lose because of: AFK players, conga lines, people using map wrong (rushing through canals for instance), etc...

I personally think a lot of the hate would also go away if W/R would not be such a primary concern. In WOT it is way bad and out of control I think. But then again, after a certain experience or many matches, I am sure that high experience players will contribute more to a game, but only over time. Still interesting to see that the guys that accuse others of being bad in terms of W/L observations, they themselves are not those experienced individuals. 

 

If I see people with over 2k games and a high WR, I figure they're quite good.

Share this post


Link to post
Share on other sites
Members
1,726 posts
15,044 battles

Unless they mainly play in divisions, in which case their win rate is anything from 10% to 25% higher than it would be when playing solo.

 

I don't know about a 25% bump for divs(in fact, that seems almost ludicrously high), but I'd believe it'll give a solid 10-15% bump for reasonably skilled players, as they usually div with the like. Nowadays I pull 65% solo, and 75-80% in a triple div, and that's pretty common for many of the players around that general "skill bracket", for lack of anything better to call it. 25% just seems like grinding an axe towards div players though.

Share this post


Link to post
Share on other sites
Beta Testers
2,196 posts

I would like to clarify something for anyone on the "windows are delicious" end of the spectrum. Shaming someone because of their stats is unacceptable. Countermanding their tactical choices or instructions in chat is a community service. 

Share this post


Link to post
Share on other sites
1,363
[HYD]
Members
7,105 posts
5,289 battles

I would like to clarify something for anyone on the "windows are delicious" end of the spectrum. Shaming someone because of their stats is unacceptable. Countermanding their tactical choices or instructions in chat is a community service. 

 

The only time i use stats to shame someone is when someone starts claiming that everything i'm doing is wrong, when it's clear they are much worse players than me. 

Share this post


Link to post
Share on other sites
10,399
[BLNCE]
Members
13,459 posts
44,054 battles

Yes, of course player skill affects win rates. Once I was at 42%, now I am over 51%. Unless you want to claim that for the last 4000 games I've been getting collapse teams in Red about 55% of the time, the only explanation is that I have improved. That improvement is shown in all my stats -- avg damage rising from 22K to 37K, K/D up from .7 to 1.2, etc. And in the way the ships I've taken up recently, like Scharnhorst or Shchors, I have solid stats, compared to the ships I played ages ago, like Furutake and Colorado.

 

There are tons of players who play mostly solo and have massively good win rates, over 60%. That can't be explained without invoking player skill, unless you want to claim that they happen to be statistical outliers who are those rare people who always get a collapse team on the other team, a statistical possibility. But then you have to explain their very high damage, kill, survival, etc numbers. For example, look at CatonBoard, no 2 player on the NA server at Warships.today. He wins 66% of his solo battles and has 77K average damage. Speaks for itself, can't be explained by freak stats. I've played against him too so I know he's good.

 

For small stat samples variation can be wide. I am a 69% winner in Benson in 79 games, but I have lower than server average damage. I suspect about 56% of that is me, and the rest is unusual luck with collapse teams that favored me. Similarly, I have a 48% win rate in ARP Kongo in 29 games, but 62K average damage -- which makes me suspect I've had bad luck with collapse teams. Ditto with Konig, where I have won only 9 of 19 games even though my average damage is a solid 59K.  

 

For large samples, you can just check the win rate and average damage -- I have 668 games in Tirpitz, with 62K average damage (above server average) and a 52% win rate (it was once 42K and 45%). Hard to dismiss either number as luck, especially knowing how much i sucked once. Similarly, at one point, playing entirely solo, I had over 400 games in Fubuki with a 59% win rate and above average damage. Hard to dismiss that as a stats outlier.

So when I check another players stats, I look at win rate and average damage. No conclusions can be made about a player with 19 games, but if they have 200 games in a ship with a 41% win rate and half the server average damage, it generally means they suck. If they have 61% win rate and below average damage, I don't worry about it. 

 

But another thing, for DDs, BBs, and CAs, it hardly matters. Any single poor player can easily be carried. But if the CV has 200 games in that CV with a 41% win rate and half the server average, you're screwed. The CV can't be carried, it has to be play well. I usually check the CV stats, I don't know why, it just depresses me....

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×