Register    Login    Forum    FAQ

Board index » General Discussion » General Chat




Post new topic Reply to topic  [ 15 posts ] 
Author Message
 Post subject: Most bashy to least bashy
 Post Posted: Fri Dec 15, 2017 1:32 am 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
In my local league I retire a race when they win a playoff. (Open schedule, I can have multiple active teams) Usually 3 playoffs each year, makes this a very long term goal.
I try to space out the different playstyle of races. I don't want to play all the elves now and not get to play 4 AG again till I am done.
I've looked at several groupings, but I decided to make a list from most bashy to least. There are plenty of races I haven't played enough to really know where they should be placed. So, I'm looking for feedback, and maybe others will find the information useful. I'm including Slann, Bretonian, and Khorne. Stunties can be hard to place because they aren't really good at anything sometimes. Mostly, I want to rank teams as they play when developed and developed optimally. Also, assume perpetual league, but redrafting every 15-20 games.

Nurgle
Chaos
Khemri
Chaos Dwarf
Dwarf
Orc
Chaos Pact
Undead
Ogre
Norse
Amazon
Khorne
Goblin
Necromantic
Underworld
Bretonian
Lizardmen
Humans
Halfling
Vampires
Dark Elf
Slann
Skaven
High Elf
Elf
Wood Elf

Anything out of place by more than a spot or two?

_________________
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Fri Dec 15, 2017 9:04 am 
Ex-Cyanide/Focus toadie

Joined: Fri Jul 24, 2009 5:55 pm
Posts: 2479
Location: Near Reading, UK
How are you defining "bashy"? Ability to cause injury? To go toe-to-toe with other teams while taking less damage? Some combination of the two?


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Fri Dec 15, 2017 10:06 am 
Super Star
Super Star

Joined: Wed May 27, 2015 10:38 am
Posts: 1017
Location: Paradise Stadium, where the pitch is green and the cheerleaders are pretty.
Dode's right that there are several definitions of bashiness, and you also have to consider that this changes as teams advance.

Chaos (and Nurgle) start off not at all bashy, all they have is ST4 on a couple guys and that is NOT enough. But after they have a few (dozen) games under their belt, they're going to have multiple killstack murderers and just tear everything up.

Dwarves start off as the lords of Bash at low team value, they start with the best bash skills (Block, Tackle, some Frenzy) and have high armor which helps prevent them from getting outnumbered ("most of the time"). And when other teams are starting to get Block, they're getting Guard and MB.
But as long as they're in a league which somehow or other tries to encourage teams of similar TV to play each other, eventually high end Dwarf teams run into those super killy Chaos teams and there's really nothing they can do about Claw. Even before that happens though, eventually other teams catch up on Block and Guard, and the Dwarves become "only" a bunch of super slow ST3 guys.

I feel like you're underestimating Lizardmen, Necromantic, even Humans (they have four guys with ST access who are reasonably easy to skill up, a Big Guy, inexpensive linemen), and overestimating Underworld, Vampires, Pact, and maybe Amazons.

Also I feel like Wood Elves are bashier than Pro Elves. Just because they start with more dodge (starting with two blodgers, and every catcher is one skill away from being another Blodger) and get a Big Guy. Then again, neither team belongs in any conversation about bashiness ;)

Skaven, on the other hand, can be sneaky-bashy (appropriate!), at least on the damage-dealing front, and I'd definitely at least put them ahead of Slann.

_________________
"It's 2+ and I have a reroll. Chill out. I've got this!"
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Fri Dec 15, 2017 10:26 am 
Veteran
Veteran
User avatar

Joined: Thu Apr 24, 2014 2:03 pm
Posts: 263
Location: Bristol
To my bashy definition is the team that his tactic is focus in do casualties to win (not if it's easy or not do it).

In that case the most bashy teams are dwarf / khemri / chaos dwarf. After that is more difficult to choose. But Nurgle and Chaos have more AG play with the pestigors, beastmen and mutations.

_________________
Image
Champion: Chimera Cup '17
Runner-Up: Exebowl '15; GertBowl '17; Exebowl '17
Stunty Cup: Exebowl '14; Bubba Bowl '17
Others: Crumb-bowl '16 (SChWing); SAWBBowl '17 (SChWing), Christmas Chainsaw Massacre '17 (Chainsaw CAS)
Wooden Spoon: Gert Bowl '14


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Fri Dec 15, 2017 10:50 am 
Legend
Legend
User avatar

Joined: Sun Nov 15, 2009 3:20 pm
Posts: 1986
Location: South Wales
Ogres are number one bashy team of all time, starting with 6 Mighty Blow :orc:

Chaos Pact come close starting with three MB.

Norse on other hand start with lots of block strength 4 and a Claw.

Necro have Claw too from the start.

Depends on the rule set your using and which teams you face too, Amazons are a bashy side, starting with 4 block and strength access.

Wood Elves have a Treeman, which puts them above Pro Elves possibly.

Its an impossible task to position them in an order like that imo. Depends on the coach using them too, for example, give Goblins to Dreamscreator and they become a bashy side, give them to me and I'll be lucky to make it to game 3 alive...

No Snotlings were hurt in the making of this post

_________________
I'm a British Freebooter, will play for any team including Undead (I have my own Apothecary). Good rates.


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Sat Dec 16, 2017 7:42 pm 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
dode74 wrote:
How are you defining "bashy"? Ability to cause injury? To go toe-to-toe with other teams while taking less damage? Some combination of the two?

"How much does the team need to remove opponent players to have a reasonable chance to win?"
More bashy -- needs to have player advantage or probably won't win
Middle -- needs to remove opponent players at least as much as their own players get removed
Least bashy -- being down players, still has high hopes

Assume BB2016 rules. Redraft after 17 games. Average of all coaching abilities.

_________________
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Sun Dec 17, 2017 12:33 am 
Ex-Cyanide/Focus toadie

Joined: Fri Jul 24, 2009 5:55 pm
Posts: 2479
Location: Near Reading, UK
CyberedElf wrote:
"How much does the team need to remove opponent players to have a reasonable chance to win?"
More bashy -- needs to have player advantage or probably won't win
Middle -- needs to remove opponent players at least as much as their own players get removed
Least bashy -- being down players, still has high hopes
Well that's something we can measure. We can compare "cas caused in a match" with "win%" for each race (I did something similar to this a good few years ago using FUMBBL data iirc) and, unsurprisingly, we find that the more cas caused by a team the higher the win%. There are other factors ofc: cas caused may not give a net casualty advantage, and net casualty advantage has a stronger correlation with win% - that may be a better measure for you. If we do we can carry out a regression to find a "cas caused" rate at which the race hits 50% (when I did this for net cas caused numbers Orcs were something like +1 net cas advantage required, for example, while WE were -2), and can compare those rates. The higher the cas caused rate required to achieve 50% the "bashier" the team.
Quote:
Assume BB2016 rules. Redraft after 17 games. Average of all coaching abilities.
And that we don't have anywhere near enough data for, certainly not under BB2016. Even with the many thousands of matches of FUMBBL data we had at the time the margins of error were pretty high.


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Sun Dec 17, 2017 5:22 am 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
Dode, at first I thought your suggestion was a great idea, then I realized it would make weak teams rank higher in the list because they need more cas to get to 50% win%. Trying to adjust for avg win% actually just takes you back to net cas I think.

What about ranking by average net casualties of only the games the team won? I guess I would be curious to see net casualty broken down by win, draw, and loss. I wonder if it is reasonable to take meaning from the difference of average net casualties of a win versus average net casualties of a loss, being bigger or smaller. A bigger difference shows that net cas is more important in determining if a team wins/loses. I guess that is a roundabout way to get to something you mentioned, that I think can be calculated more easily. Which races are net cas more strongly correlated to win/draw/loss?
(I left my train of thought, but I'm really only interested in the last question now.)
I accept losing the specific environment to gain hard numbers.

_________________
Image


Last edited by CyberedElf on Sun Dec 17, 2017 5:06 pm, edited 1 time in total.

Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Sun Dec 17, 2017 11:41 am 
Super Star
Super Star

Joined: Wed May 27, 2015 10:38 am
Posts: 1017
Location: Paradise Stadium, where the pitch is green and the cheerleaders are pretty.
if you have individual match stats and only racial performance, what you can do is figure out how much each race's performance W/L improves when their net casualties improves.

So for each race, chart out what their W% is when they're at -2, -1, 0, +1, +2 net casualties. That boundary is arbitrary, but I suspect that the number of games outside that range is much smaller than the number of games when both teams are within 2 CAS of each other. It doesn't tell us much to know that teams will win a lot when they're at +5 casualties, and so on.

Wherever you get your data from will also seriously affect your results. You want data from leagues with many games, and which also matches your scenario. Your best bet would be perpetual leagues on FumBBL. I'm not sure if they have the option to, say, "show me all Orc games and dump them into a .csv file" or something. That kind of info IS available for NAF tournaments, but obviously that's a seriously different environment.

Anyway, the steeper the lines, the more that race relies on bash for its success. I'm guessing the lines won't be totally linear per race though, even if you could get a perfect sample size. Khemri suck, for example, and while it gets easier to win games when you remove an opponent or two, it really hurts to play Khemri outnumbered. But in general, the higher each marginal casualty increases a race's win percentage, the bashier it is by your definition.

Sounds like a lot of work though.

_________________
"It's 2+ and I have a reroll. Chill out. I've got this!"
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Sun Dec 31, 2017 4:50 am 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
Still working on getting raw data from FUMBBL. In the meantime I found a summery of FUMBBL data; BB2016, Blackbox. I sorted by net casualties.

    Dwarf
    Orc
    Nurgle
    Chaos Dwarf
    Khemri
    Chaos
    Undead
    Chaos Pact
    Necromantic
    Norse
    Human
    Lizardmen
    Amazon
    Dark Elf
    Underworld
    Slann
    High Elf
    Ogre
    Skaven
    Vampire
    Elf
    Wood Elf
    Halfling
    Goblin

Not yet the definition I want to use, but something other than just going by feel.
http://fumbbldata.azurewebsites.net/stats.html

_________________
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Tue Jan 23, 2018 4:17 am 
Super Star
Super Star

Joined: Tue Jan 17, 2006 7:55 pm
Posts: 859
Mmm, I think for the last table to be meaningful there needs to be some sort of analysis on whom is killed on either team - Ogres rarely die, cause a lot of damage. Snots in reverse virtually never cause a CAS (at least not from blocking), but will often die. At 20k to replace, they are worth a sixth of an Ogre. Apples and Oranges.


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Tue Jan 23, 2018 4:12 pm 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
I think I got FUMBBL data, but there have been problems reading it into Google, Excel, or R. I will still try to answer which races have the highest correlation between net casualties and winning, but I found another question that I think is more useful. Net casualties vs. winning for a race ignores any component of the same thing for their opponent. For example, Nurgle may have a high correlation, but it is diluted when they play a team like Wood Elves that have a poor correlation. Maybe a better way to compare style of play for a race is correlation between net casualties and touchdowns for that team. If I ever get this project done, I will probably post both results, but I think the second better matches my question of relative "bashiness" of each race.

Aliboon wrote:
Mmm, I think for the last table to be meaningful there needs to be some sort of analysis on whom is killed on either team - Ogres rarely die, cause a lot of damage. Snots in reverse virtually never cause a CAS (at least not from blocking), but will often die. At 20k to replace, they are worth a sixth of an Ogre. Apples and Oranges.

The last list, as stated, is presented because it was the data readily available. What question are you trying to ask such that analysis will give you an answer, please be very specific? I don't know of any large dataset of Blood Bowl matches that includes which player was injured.

_________________
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Tue Jan 23, 2018 5:44 pm 
Emerging Star
Emerging Star
User avatar

Joined: Wed Oct 15, 2014 6:11 pm
Posts: 495
Does number of CAS take into account Fouling? I'm not sure if Bashy teams are more or less prone to foul an opponent. I think it largely depends on whether the player Coach has expendable players to foul with and possibly a numbers advantage on the opposition at the time. Just food for thought.

_________________
Bloggin it: https://doubledowndice.wordpress.com


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Wed Jan 24, 2018 4:43 am 
Veteran
Veteran

Joined: Fri May 31, 2013 1:52 am
Posts: 225
@faust_33 The previous list was based on data that counted opponent casualties suffered. So yes, it included fouling and crowdsurfs, but also failed GFI and dodge.

I finally got the data and stats help. The results did not all come out as expected. You can see (crappy) graphs, exact values and confidence intervals here. I only had data for three months of FUMBBL. I only used data from the Blackbox division. As I understand it, you can pick your opponents in Ranked, but Blackbox has more random pairings.
Based on which races have the highest correlation between net casualties and touchdowns they scored the list from high to low is:
Chaos Dwarf
Chaos
Dwarf
Norse
Amazon
Chaos Pact
Undead
Human
Khemri
Orc
Underworld
Nurgle
Necromantic
Lizardmen
Goblin
Halfling
Ogre
Dark Elf
Elf
Slann
Wood Elf
Skaven
High Elf
Vampire

To me, Khemri, Orc, Nurgle and High Elf were surprisingly low while some of the hybrids were higher than I expected. I also ran the correlations for net casualties vs winning, casualties suffered vs touchdowns, and casualties suffered vs winning. The outcomes of those three analysis can also be found at the above link. If you don't go look, the short version is that the other comparisons did not even come close to expected order of "bashiness." If you look at the graph, you can see that all elves, Slann and Skaven (and Vampires) have a substantially lower correlation than all of the other races. While not surprising, the drop off was significant.
Of significant note, the data I had was only casualties suffered by opponent not casualties caused. I suspect this is why Vampires had the lowest correlation in all four analysis.
I just realized I ran calculations on casualties suffered by that race not casualties suffered by the opponent. My misinterpreting the data labels should not change net casualty analysis, but I need to run two of the calculations again.

_________________
Image


Top 
 Profile  
 
 Post subject: Re: Most bashy to least bashy
 Post Posted: Wed Jan 24, 2018 10:15 am 
Super Star
Super Star

Joined: Wed May 27, 2015 10:38 am
Posts: 1017
Location: Paradise Stadium, where the pitch is green and the cheerleaders are pretty.
I guess that kind of goes to show that Net Casualties are pretty irrelevant when evaluating how "bashy" a team is.

Khemri's a great example. It's easy enough to get some CAS against those AV7 skeletons. It's very common with Khemri to get out-casualtied.
But the OP's definition of bashiness was the amount the team relies on outnumbering opponents in order to win games. That's Khemri to a ... K. It's hard to score with Khemri at the best of times, and very hard when even up or outnumbered. Khemri do take their share of Casualties, they rely on a deep bench of cheap skeletons, Thick Skull (meaning they lose less KO's per armor break), Regeneration (Half their CAS come back), and fouling to get a numbers advantage.
None of that is visible when you look at "Net Casualties".

It's possible to measure that empirically but not with the stats available. If we had a couple hundred volunteers, they could go through replays of games and look at every team's offense.
There are three possible outcomes to an offense: Touchdown, Turnover (in the Football sense, meaning losing control of the ball), or "Half ends with no score".
Any time of those results occurred, our hypothetical volunteers would record how many players the offense has minus how many the defense has. So "+2" means a team scored while outnumbering their opponent by two, "-1" means they scored despite being down a player, and so on.
After recording this for hundreds or thousands of games, you'd be able to see how each race's offense performs when even, outnumbered, and having more players. So we'd have data like, "At even strength, 40% of Khemri drives end in touchdowns, 30% end with no result, 30% turn the ball over. At +1 Player, 55% of Khemri drives end in touchdowns, 25% end with no result, 20% turn the ball over".
Plot out a team's offensive outcomes for each situation where there's a decent amount of data. The more it improves with each additional player advantage, the "Bashier" a team is by the OP's definition.

We don't have lots of volunteers to go through replays for us, so probably the best way to get a usable answer for the OP is subjective guesswork based on experience playing those teams. And my subjective experience says that any "Bashiness" ranking which ranks Humans as bashier than Orcs and Khemri, Ogres next to Dark Elves and Pro Elves, is probably measuring the wrong things ;)

_________________
"It's 2+ and I have a reroll. Chill out. I've got this!"
Image


Top 
 Profile  
 
Display posts from previous:  Sort by  
 
Post new topic Reply to topic  [ 15 posts ] 

Board index » General Discussion » General Chat


Who is online

Users browsing this forum: No registered users

 
 

 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to: