Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

I'm having a bit of fun cracking open the BCP data. Eventually I'll get my cookie to stick and I can analyze data by sub faction so stay tuned ( hopefully ).

Anyway - I ran a years worth of 28+ person tournaments ( didn't want to pull too much data ) and it gives me a unique ID for each player so I can make sure everyone is distinct. It returned a total of 9,887 players.

In the file it shows each army they played during this period and the times they played it. Some are listed as their subfaction so 'Freebooterz' instead of 'Orks'. This is not a switch, of course, but that would take too much coding right now. There are also a lot of people who soup. Some might play Thousand Sons and then Tzeentch. Or Imperium and then some flavor or marines and then Knights. These people didn't likely switch either.

All of that makes for some hard to parse data, but what about the raw figures?

2 players switched 7 times
2 switched 6 times
13 switched 5 times
38 switched 4 times
134 switched 3 times
423 switched twice
1,474 switched once
7,799 never switched at all -- ( there's a fair number of people who played only once so they've had no chance to switch, but they either just don't play often and in the case are extremely unlikely to switch or they just fell out of the gaming side of the hobby )

A lot of the people who switched only once went from things like Salamanders to AA, Harlequins to Aeldari, AM to Imperium, AM to Cadian Shock, etc.

Respect to the World Eaters player who has done 10 tournaments.

You can see all the data here:
https://docs.google.com/spreadsheets/d/1F8Qom_A8N5e5aVt5pMLe3oqDDV648WSC5nCpxpzFq3Q/edit?usp=sharing

Where I was from the tournament players hosted a buy/sell group on facebook and were constantly getting rid of armies on edition changes and rebalance passes.

But I use the word "tournament players" to mean the hyper competitive players. They made up about 40% of the population.

The rest of the people that went to tournaments were not hyper competitive but went to tournaments regularly and even went to like adepticon to play.

I think there is a distinction between the two.

 auticus wrote:
Where I was from the tournament players hosted a buy/sell group on facebook and were constantly getting rid of armies on edition changes and rebalance passes.

But I use the word "tournament players" to mean the hyper competitive players. They made up about 40% of the population.

The rest of the people that went to tournaments were not hyper competitive but went to tournaments regularly and even went to like adepticon to play.

I think there is a distinction between the two.

They definitely exist. It takes a while to run this so I'll parse a year at a time and grab smaller tournaments too and see what changes. Should be interesting nonetheless.

Very interesting data, Daedalus. If I were to make two buckets, I'd say the no-switchers and single-switchers should be lumped together, and then everything above that would be the other bucket, and the first bucket would be the more casual folks, and the second bucket would be more hardcore tourney players. So even when we restrict ourselves to ostensibly competitive/tournament players, only about 5% of those are really jumping around to min-max their competitive chances. I think that's a pretty healthy number.

Thanks for doing this!

 Gene St. Ealer wrote:
Very interesting data, Daedalus. If I were to make two buckets, I'd say the no-switchers and single-switchers should be lumped together, and then everything above that would be the other bucket, and the first bucket would be the more casual folks, and the second bucket would be more hardcore tourney players. So even when we restrict ourselves to ostensibly competitive/tournament players, only about 5% of those are really jumping around to min-max their competitive chances. I think that's a pretty healthy number.

Thanks for doing this!

I think the first thing I would do is remove all the "only played once" people

One thing to remember is that just because they played one faction and then another doesn't mean they got rid of the old one. Most likely they still have their old miniatures. Most people eventually have multiple armies.

Breton wrote:
I think the first thing I would do is remove all the "only played once" people

I only ran a year and bigger tournaments, which is fairly limited in scope. I am curious to see if we get more data on them as I open it up.

 Grey Templar wrote:
One thing to remember is that just because they played one faction and then another doesn't mean they got rid of the old one. Most likely they still have their old miniatures. Most people eventually have multiple armies.

I think the issue is a lot of people have a main army they can play at tournaments - and then a lot of stuff they can mess about with, or play smaller games with, but maybe don't have a viable 2k points (or know a lot of that 2k points is bad, and going to a tournament to be smashed is... well, I'm not a fan anyway).

Technically, now that you have them tracked by player IDs, you can run a really important analysis.

Faction win rates normalized by player's average score.

Spoletta wrote:
Technically, now that you have them tracked by player IDs, you can run a really important analysis.

Faction win rates normalized by player's average score.

So like where better players bring a faction's performance up?

We always have issues with win rates of factions because as soon as a faction isn't in the top, the competitive players switch faction and create an avalance effect.

We never understand if factions are bad or the players of that faction are simply less of the competitive kind.

If instead of using a W/L metric for the results of a faction, you use overperformed/underperformed compared to that player average scores, you get a much more accurate view of the actual balance of the factions.

Also some players might own multiple armies and simply take a different army to each tournament just for the variety. So some of those list switchers could just be casual players but prolific painters.

This is very interesting - thanks for doing this! Concur that removing single instance players will give you more clarity.

My observation of my local meta is that a majority of us have a primary army. They will tweak it between tourneys but otherwise stay loyal to a faction. About a quarter switch armies from time to time, but they tend to already have those armies. A very small handful go out and wholesale buy new armies based on the new hotness.

Having said that, Indomitus led to an increase of players with Necrons. Not faction jumping, but folks trying something new with the new models they got.

Ok so I ran 3 years and 16 man tournaments. There are 24,892 players.

Never switchers
About 13,600 played just once in three years. This doesn't mean they don't play more, but simply either attend smaller tournaments or just go casual ( or quit! ).
1,800 played twice
659 played thrice
256 played four
380 played five or more times

4087 switched once
4011 switched twice
2119 switched more than twice

The data is up on the link above on a new tab.

Of course the amount of switching is less than perceived, because of players like this guy:

4 - Militarum Tempestus 4 - Astra Militarum 1 - Adeptus Titanticus 5 - Catachan Jungle Fighters 1 - Imperium

Anyway it' a fun little data set to peruse. Enjoy!




Automatically Appended Next Post:

TangoTwoBravo wrote:
Having said that, Indomitus led to an increase of players with Necrons. Not faction jumping, but folks trying something new with the new models they got.

Yea you can definitely see the people who collected Necrons - some will have a sudden spike with them in the middle of their "main" army.


Automatically Appended Next Post:

Spoletta wrote:
We always have issues with win rates of factions because as soon as a faction isn't in the top, the competitive players switch faction and create an avalance effect.

We never understand if factions are bad or the players of that faction are simply less of the competitive kind.

If instead of using a W/L metric for the results of a faction, you use overperformed/underperformed compared to that player average scores, you get a much more accurate view of the actual balance of the factions.

That might be difficult to do still.

There isn't anything that indicates to me who is experienced. So it's hard for me to say Player A who is using Drukhari is just winning, because the army is busted or winning, because they're a good player.

Hmm...



Get win rates for every Admech player. Then overlay the number of times they switch. So, 1 or 2 switches would be a group, 3 to 4, and 5+. Then we can see if there's any noticeable different between static players and those that jump around more. That probably won't get us exactly what you're looking for, but it should at least be interesting.

Well...bleh...there's a time value to the switching. So if they switch away when did they do it and how did it affect things. More noodling to do...

I would do it like this:

Filter away any player who hasn't played at least 3 factions. We are interested in how much switching faction alters your win chances, so those who played one or two factions are not relevant to this analysis.

Out of the remaining players, calculate for each own its own W/L ratio, combining all his results. This should give a value between 0 and 1. We will call this the APWR (Average Player Win Ratio).

With these results, make an APWR reference table.

Then, you parse just the most recent results, like the last 6 weeks.

Apply the following algorithm:

Each win is worth 1-APWR.
Each loss is worth 0 - APWR

Calculate the average of those results for the factions.
If the result is higher than 0, then playing that faction does indeed increase your chances to win.
If the result is lower than 0, then playing that faction comes as an handicap.

0 is the ideal value.

This way, you remove any influence of the actual skill of the players of a faction.
We take only the last 6 weeks because we are interested in the current meta, but we take the APWR considering the full 9th edition, because we assume that a good player stays a good player.

Gotcha. I am going to have to map subfactions into one to avoid noise, but I will get to work.

Breton wrote:

 Gene St. Ealer wrote:
Very interesting data, Daedalus. If I were to make two buckets, I'd say the no-switchers and single-switchers should be lumped together, and then everything above that would be the other bucket, and the first bucket would be the more casual folks, and the second bucket would be more hardcore tourney players. So even when we restrict ourselves to ostensibly competitive/tournament players, only about 5% of those are really jumping around to min-max their competitive chances. I think that's a pretty healthy number.

Thanks for doing this!

I think the first thing I would do is remove all the "only played once" people

But then you've thrown out a lot of 'only one army/never switched' people. That's a really hard bias to inflict on the data set- hard enough to render it meaningless.

Of course, this assumes that if they don't play in further tournaments (and these specific tournaments in particular) they fell out of the hobby, so we're already knee-deep in problematic uses of the data.

7,799 never switched at all -- ( there's a fair number of people who played only once so they've had no chance to switch, but they either just don't play often and in the case are extremely unlikely to switch or they just fell out of the gaming side of the hobby )

Voss wrote:

Breton wrote:

 Gene St. Ealer wrote:
Very interesting data, Daedalus. If I were to make two buckets, I'd say the no-switchers and single-switchers should be lumped together, and then everything above that would be the other bucket, and the first bucket would be the more casual folks, and the second bucket would be more hardcore tourney players. So even when we restrict ourselves to ostensibly competitive/tournament players, only about 5% of those are really jumping around to min-max their competitive chances. I think that's a pretty healthy number.

Thanks for doing this!

I think the first thing I would do is remove all the "only played once" people

But then you've thrown out a lot of 'only one army/never switched' people. That's a really hard bias to inflict on the data set- hard enough to render it meaningless.

Of course, this assumes that if they don't play in further tournaments (and these specific tournaments in particular) they fell out of the hobby, so we're already knee-deep in problematic uses of the data.

7,799 never switched at all -- ( there's a fair number of people who played only once so they've had no chance to switch, but they either just don't play often and in the case are extremely unlikely to switch or they just fell out of the gaming side of the hobby )

I'm of two minds.

These played once people can be hidden in smaller tournaments. They're much less likely to be swayed into meta chasing and they capably collected 2,000 points for an army.

Switching armies is a normal thing. The questions become -- how often, to what armies, and for what reasons? If only 6 to 7% of the tournament base is switching armies regularly. Another 15% might have switched once for cause. And that's kind of it.

What will be interesting to see is tracking people when the next big book hits and checking those same people to see if any of them switch.

Spoletta wrote:

Out of the remaining players, calculate for each own its own W/L ratio, combining all his results. This should give a value between 0 and 1. We will call this the APWR (Average Player Win Ratio).

With a sample size of 3-4 games for the majority of players that ratio doesn't seem very informative.

Yeh. You could get anything from 0 to 100 for same player with that sample size

 Altruizine wrote:

Spoletta wrote:

Out of the remaining players, calculate for each own its own W/L ratio, combining all his results. This should give a value between 0 and 1. We will call this the APWR (Average Player Win Ratio).

With a sample size of 3-4 games for the majority of players that ratio doesn't seem very informative.

Players that were not filtered away by "Played only one or two factions" have for sure many many more than 3 or 4 games.

With each faction though? One can have wr between 0 and 100 for section of 3-4 games easily. Not going to give much info anyway even comparing 3-4 games per faction.

Not really sure you can glean from the data. If someone has say 2 armies they can run, them switching appears the same as someone who trades in their collection for another 2k point force.