There used to be a site that did just that. I can't remember the name but I'm sure someone will.

Of course in the end it wasn't very reliable because each and every list had gone through several versions, but I think it would be great to start again from scratch now that most lists require only minor modifications.

Wow, head off to a tournament for the weekend and I get to miss all this.

Whilst I agree that Battle Reports are not the be all and end all of testing I would like to add a cautionary note. Several years ago there was a constant claim about the ineffectiveness of an unit. For the life of me I can't remember who or what the problem was but several others couldn't understand why this tester was getting these results.

On questioning and I think finally with a Batrep it was shown that they were using less terraine and a 4'x4' table and not the recommended 6'x4'. These changes made a huge difference to the results which then became clearer to those of us not understanding why the inital comments were made.

I'm not suggesting that people are testing inappropiately, just that by using batreps those of us not in the same room can understand what is being highlighted.

Note that Theoryhammer is great, but my all powerfull LatD list lost twice over the weekend in circumstances that I hadn't expected.

Hope that I haven't embarrassed anyone with my example above and sorry if I have.

_________________
_________________
www.epic-uk.co.uk - home of the UK Epic tournament scene
NetEA NetERC Xenos Lists Chair
NetEA Ork + Feral Ork + Speed Freak Champion

Win/loss ratios may not give a very accurate representation of army list abilities.

Last year I played many more games than this year. Most of those games were against an opponent that played Marines exclusively. He was one of the players who re-introduced me to the game after a 15 year break. After the first 2 or 3 intro games, he didn't beat me again before he had to move away (military posting). Thats maybe 10-15 games where I used Orks, Tau, I.G. and Marines and didn't lose.

That win/loss ratio would be of little value. Win/loss ratio's depict player abilities far more than list abilities.

If stats are going to used then I think Hena's previous post would be more beneficial. More detail is needed (such as that which you get in a batrep) to be able to make informed decisions.

_________________
Just call me Steve.

NetEA Rules Chair
NetEA FAQ

Want to play Iron Warriors in Epic Armageddon? Click HERE
Some of my Armies.
My Hobby site.

Onyx, true, in small numbers and for a small number of player. However, as the number of players and reults increases the ability of the players averages out.

Reports by 1 marine player will be biased to their abilities. Reports by 10 marine players will be reaonably average.

I'll keep track of how many different players each army has, and use that when calculating the margin of error.

_________________
http://www.troublemakergames.co.uk/
Epic: Hive Development Thread

Hena: While that would be ideal, my web programming skills barely go beyond javascript these days, I'm totally out of practice.

I think for starters a straight results thread compiled in excel will do, and if it proves successful enough I'll look at coming up with something better.

_________________
http://www.troublemakergames.co.uk/
Epic: Hive Development Thread

I really think we need more data than just "Armies" and "Who Wins".

Point values played and victory conditions achieved might be good as well, and even a picture of the initial battlefield.

It's a valiant attempt, but it's so information-lite as to be next to useless.  Sorry, Zombocom.

_________________
"EPIC: Total War" Lead Developer

Now living in Boston... any EPIC players want to meet up?

Chroma: Drawing statistics from such a detailed picture is next to impossible, particularly with the number of games likely to be reported. A straight win/loss ratio is useful for one thing at least, direct comparisons of overall army list balance, an that's all I intend to use it for. This is all the info we got from the UK Tournaments, and I found it extremely useful.

As I previously said, if this proves successful then I'll look into coding up a more complex version

Once we have 25 games reported for a specific army the margin of error drops below 10%, at which point I'll start posting current percentages and margins of error.

_________________
http://www.troublemakergames.co.uk/
Epic: Hive Development Thread

Without knowing the winning conditions of the games played, the "win/loss" ratio is relatively meaningless, except as the most coarse of measures.

A 5-0 win is *not* that same thing as winning by a difference of 150 VPs, but look the same on the "win/loss" table; in my opinion, you'd need a "weighted value" to types of wins, not just straight wins, to get *useful* data out of this endeavour.

Army composition is also a very large factor in determining balance.  A "win/loss" record does nothing to examine this, and it's a *very* important part of army balance.  Say, over the next month I play ten "Murder of Monolith" games and I dominate, winning 8 out of 10 games against various armies... while someone else is playing a "Very Very Venator" army, because they like Flayed Ones, and gets trounced 8 out of 10 games against mixed armies...  without examining composition, those results come out "balanced" in analysis and are as useless as a single battle report that says, "See, they can lose!"

_________________
"EPIC: Total War" Lead Developer

Now living in Boston... any EPIC players want to meet up?

Could we just take a step back a second here and check that we all have the same aims in mind? So what are we trying to determine when having these sort of discussions on the board? For example, are we confirming that a given list is 'balanced', or that a particular unit is 'unbalanced'; against whom, and how is that determined etc? Indeed, how do we even determine whether a given unit should be represented (fluff, equivalence in 40K, 'to fill a gap' etc).

IMHO a large part of the problem would seem to be a lack of agreement on the criteria to be used - or put another way, how are we to judge objectively whether a particular claim is valid or not? For what it is worth, among other things I expect the army champs and NetEC to provide the relevant guidelines and preferably to explain their thought process from time to time so that we can all work constructively towards common goals. Equally I hope this might also reduce the currently tiresome level of general bickering and circular arguments.

With regards to Batreps, while I agree we need some extra details along the lines provided by Hena, we also need to consider what we are trying to achieve as well as understanding what is influencing the player’s decisions. So is this a competition battle with army and battlefield decisions chosen accordingly, or was the battle set up to prove a particular point etc?

Finally, statistics may be usefull as an illustration to a point, but as Chroma points out they need to be treated with great care and put in context. It may be perfectly reasonable for the average person to travel at 70mph on a motorway on a sunny day but quite a different matter to drive at 70mph on a small country road at night in fog and rain, and equally different if the driver is an experienced.

_________________
"Play up and play the game"

Vitai lampada
Sir Hemry Newbolt

Tim they were playing from short edge to short edge on a 6x4 table and as a result found CC troops rather poor.

Oh and several people a while back expressed amazement at terrain density on tables with 12 pieces, they using a fair bit less

_________________
If using E-Bay use this link to support Tac Com!
'Abolish red trousers?! Never! Red trousers are France!' – Eugene Etienne, War Minister, 1913
"Gentlemen, we may not make history tomorrow, but we shall certainly change the geography."
General Plumer, 191x

It is the same when determining who won the game.

Straight win/loss is fairer than, say, average points scored, because some armies tend to win big/lose big, (AMTL for example) whereas others usually have tighter games. The margin of error would be thrown off due to this, but it wouldn't be calculatable.

Again, this is just a simple test intended to get win/loss ratios, nothing more.


without examining composition, those results come out "balanced" in analysis and are as useless as a single battle report that says, "See, they can lose!"

_________________
http://www.troublemakergames.co.uk/
Epic: Hive Development Thread

Actually, since the data you're looking for would be a subset of a larger gathering of information, more could be made out of anything analyzed and one could still look at that simple win/loss ratio while also allowing for a richer analysis.

It's easier to get everyone to submit that data now, rather than ask for more information later, when they may have forgotten it.

_________________
"EPIC: Total War" Lead Developer

Now living in Boston... any EPIC players want to meet up?

Fair enough, I'll edit the thread to ask for scenario points scored, as that is at least averagable.

Psrticular build types aren't really possible to analyse in this way, so I'll leave them as optional for now.

_________________
http://www.troublemakergames.co.uk/
Epic: Hive Development Thread