Five Tribes Five Tribes

Forums

Search
Forums » T2R Competitive Play - English » On the importance of starting
Show: Today's Posts 
  
AuthorTopic
AAA_Angel6
Senior Member
T2R Nation Cup 2007 Winner

Posts: 1055
Registered:
November 2004
On the importance of starting Fri, 24 February 2012 11:32
Abstract

My guess is that there is a less than 10% advantage for the starter of a T2R game, which would mean that the starter has an overall chance of 55% to 60% to win a game. Imho, many players believe the advantage is much greater than that Smug

In this threat I try to approve my guess by interpreting an experiment of mine that lasted for the last 15 or so months.

After this (and after the end of the CL 2012) I plan to reconfirm (or disprove) my guess by evaluating the 3 bygone Champions Leages, because they are the only tourneys that provide the starting information to every game.

My instincts tell me that starting is much more important against top players than against weak players, but I'm not sure if I can evaluate this Question


Introduction

Probably all tournament players agree that the starter has a significant advantage in T2R Smile

As far as I remember, thekid used to lament about his starting unluckiness from 2005 on. Back then I just thought: ok, thekid plays badly without starting, but it doesn't really matter THAT much Wink

In the end of 2010 I decided I'd like to evaluate the "exact" advantages of starting. The idea that I had was to open a 2nd account ( Embarassed ) and with this account I would never start a game and finally compare the maximum ranking with that of my main account.

Pavonine, I wanted to maximize my non-starting ranking, so my 2nd account was secret which might have given me some advantages (I only postet it in the "Declaration of Honor" threat, but who should read this as I only edited my old post from 2008 Razz ).
Only before the start of NC 2011, I informed my team and the TDs because of the strict NC rules Wink
      
The Good Guy
Junior Member

Posts: 1
Registered:
December 2010
Re:On the importance of starting Fri, 24 February 2012 12:47
Parameters of the experiment

410 games were played, all the same:
-) 2 players, US map (no 1910);
-) the opponent was allowed to start;
-) closed games, nobody could watch, no cheating possible;
-) game name was "1300+", I played everybody accordingly.

I was asked about my "identity" a couple of times. Trying not to lie but not telling about Angel, I played possum and told my (real) 2nd name Matthias. Sometimes it was real funny, e.g. when I asked Mudda about his name and he wrote Elfriede Very Happy

Thx for many opponents to point out to me that I untentionally let them start Thumbs Up

There were good and bad times Wink
Some average+ opponents like brainaid and ClayMaster gave me the impression I played real stupid Evil or Very Mad
Some tops like RMarkes and Tenniskasi let me believe that starting is unbeatable Twisted Evil
I had a very nice 7:5 vs Rufuz with gave me much confidence after a 10:14 vs foudecoasters and a 6:10 vs foss Very Happy
And some players like Hunsrückbahn just couldn't accept that they can't beat me while starting Very Happy

Right now I feel that I'm already above my top limit with overall 1706. At #2 or #3 I fancied going for #1, but in fact I'm so far away and much too dependent of the rival's dropping scores.
I might be able to win a few more games, but on the long-run I feel like settling for something like 1680 where I belong.

Experiment closed, even though I'm not sure how to evaluate this at the moment Rolling Eyes

[Updated on: Fri, 24 February 2012 13:42]

      
AAA_Angel6
Senior Member
T2R Nation Cup 2007 Winner

Posts: 1055
Registered:
November 2004
Re:On the importance of starting Fri, 24 February 2012 13:08
Comparabiltity

Are the rankings of Angel6 and The Good Guy comparable?

PROS:
-) Opponent structure: any 1300+, that's what I play with both accounts. Probably also a similar amount of tops (not very many with both accounts).

CONS:
-) # of games: 1510 vs 410, I think that my average concentration level was higher with The Good Guy.
-) Multis: Angel played 26 multis with not much success - only a minor factor imho.
-) Maps: 244 EU, 52 Swizz, 45 1910 variants, only a minor factor even though I dropped significantly in EU.
-) Most games with Angel open, so I might have been cheated from time to time, sometimes it feels like the opponent knows everything Wink

Angel's top ranking during 2011 was at least 1749, a difference of 43 points. Difficult to evaluate the "advantages" of The Good Guy, but I'll say that the difference between 50% starting and 0% starting for me is 60 ELO points.

Now I need time to get the math right, to be able to say "not starting is costing me the win in ..% of my games" Smile
      
ACP Miguel
Senior Member
T2R Swiss Map Championship 2011 Winner

User Pages
Posts: 591
Registered:
October 2006
Re:On the importance of starting Fri, 24 February 2012 15:13
well not my field of expertise but some biased factors there.

# the opp of course, u cant really be sure the level of opp as u place +1300, playing 410 gmaes u could play very few top players, as for the Angel account u probably get more tops, specially if u are higher ranked people tend to take an opp with more points (at least i do...)

# if u are trying to isolate a variable (starting) then in order to make some stats analysis u would have to have similar samples of other variables (per example: the opp u play, the ticks u had while non starting).
so in theory to analyze that u would have to play a significant amount of games vs same opp in each branch of study, and playing the same ticks, and having the same coloursm locos distribution - other than that u are just assuming that the result u had was caused by variable start and letting all others apart.

# the fact the users dont know who they are facing. Diff strategies apply when u are playing someone u know (more blocking, starting earlier to take routes, etc)

but i would like to hear the stats guys

cheers
miguel
      
Sysyphus - Pommard
Senior Member
T2R All Around Tournament 2011 Winner

User Pages
Posts: 2493
Registered:
December 2007
Re:On the importance of starting Fri, 24 February 2012 16:13
Another factor important to me is your way of playing.

If you are a master in designing longest, the start would less matter if you play someone who has less skills for that (I.e me Very Happy ).

On the contrary, at equivalent skills on designing longest, start surely matters as much as managing your colours the best way.

And on the opposite side, someone who likes to go out fast would enjoy more the start because it fits better with his style of playing. The extra turn given usually gives my opponent a good opportunity to lay a bigger track at the end than a smaller track needed to obtain longest.
But, it's up to me to correct that flaw Cool .

Start is way less important than tickets in multis but does matter if you target longest and big tracks. It'd be interesting to see the influence of that factor for players like will or womble.
Not having the start in multis usually gives an extra chance for opponents to fiend more tix, often causing damages.
Playing in a multi without anyone able to finish with 44 cards is a true paradise for fienders.

But I'd tend to agree with angel : a good management of colours and an expertise in designing longest can indeed make up for the absence of start.
It would be interesting to have the light of maunaloa on his stats.

[Updated on: Fri, 24 February 2012 16:17]

      
AAA_dea1
Senior Member
T2R Nation Cup 2007 Winner

Posts: 1957
Registered:
September 2005
Re:On the importance of starting Fri, 24 February 2012 18:34
TRUE Miguel schrieb am Fri, 24 February 2012 15:13

# the fact the users dont know who they are facing. Diff strategies apply when u are playing someone u know (more blocking, starting earlier to take routes, etc)



Actually, that could be looked at from now on.
Keep playing GoodGuy for some more games, Angel, and let's see if there really is a difference if people know, who you are - would you?

      
AAA_Lucullupus
Senior Member

User Pages
Posts: 753
Registered:
July 2008
Re:On the importance of starting Sat, 25 February 2012 13:12
Therefore it's time for a new account! Let's say the "even better guy"? Laughing

[Updated on: Sat, 25 February 2012 13:12]

      
AT Tyrana
Senior Member

Posts: 184
Registered:
August 2007
Re:On the importance of starting Sat, 25 February 2012 17:44
or "even better girl" Rolling Eyes Very Happy

We can put this to the test when we play CL Laughing
      
Sysyphus - Pommard
Senior Member
T2R All Around Tournament 2011 Winner

User Pages
Posts: 2493
Registered:
December 2007
Re:On the importance of starting Sat, 25 February 2012 18:02
DD dea1 wrote on Fri, 24 February 2012 10:34



Actually, that could be looked at from now on.
Keep playing GoodGuy for some more games, Angel, and let's see if there really is a difference if people know, who you are - would you?




Again, it depends on your style of play and the way your opponent influences your style of play.

Strength of Angel (well this is how i see him) is too keep a very good balance between flexibility and efficiency all through the game, giving him multiple options after every move of his opponent, whoever his opponent is.
So, for me, knowing I play him or him as unknown second account wouldn't change much cause it influences less my way

It'd change more if you play against someone with a more marked style of play.

[Updated on: Sat, 25 February 2012 18:11]

      
ATN blubes
Senior Member

Posts: 143
Registered:
April 2008
Re:On the importance of starting Sat, 25 February 2012 19:49
hi folks,

ancient stats from maunaloa


| 1st | 65.8% ( 73/ 111) |
| 2nd | 35.5% ( 43/ 121) |

nuf said

b
      
ATN Drake
Senior Member

Posts: 819
Registered:
March 2006
Re:On the importance of starting Sun, 26 February 2012 00:02
I tracked my last 200 or so 1400+ 2P matches.

Started 138 times
Won 93 (Got longest 67 times, tied long 4, not long 22)
Lost 42 (Got longest 9 times, tied long 4, not long 29)
Tie 3 (Got longest 2 times, tied long 0, not long 1)

Non-Starter 132 times
Won 58 (Got longest 41 times, tied long 5, not long 12)
Lost 67 (Got longest 15 times, tied long 6, not long 46)
Tie 7 (Got longest 2 times, tied long 1, not long 4)

Think I would rather go first and get longest!
      
UEG xbomanx
Senior Member
T2R Single Player World Championship 2010 Winner

User Pages
Posts: 339
Registered:
May 2008
Re:On the importance of starting Sun, 26 February 2012 01:31
DrakeStorm schrieb am Sun, 26 February 2012 00:02

I tracked my last 200 or so 1400+ 2P matches.

Started 138 times
Won 93 (Got longest 67 times, tied long 4, not long 22)
Lost 42 (Got longest 9 times, tied long 4, not long 29)
Tie 3 (Got longest 2 times, tied long 0, not long 1)

Non-Starter 132 times
Won 58 (Got longest 41 times, tied long 5, not long 12)
Lost 67 (Got longest 15 times, tied long 6, not long 46)
Tie 7 (Got longest 2 times, tied long 1, not long 4)

Think I would rather go first and get longest!


Nice report Drake, but how can you track your last 200 matches??? I want it too.

Greetings
xbomanx
Very Happy
      
ATN Drake
Senior Member

Posts: 819
Registered:
March 2006
Re:On the importance of starting Sun, 26 February 2012 01:44
G70 xbomanx wrote on Sat, 25 February 2012 16:31



Nice report Drake, but how can you track your last 200 matches??? I want it too.

Greetings
xbomanx
Very Happy


It is tedious. Log results into Excel after every match.
      
SOJA Patterson
Senior Member
T2R 2010 World Champion

User Pages
Posts: 220
Registered:
January 2006
Re:On the importance of starting Sun, 26 February 2012 15:29
For the number lovers, some extra bits to crunch....

I conducted a similar exercise as Drake, but I did it in early 2011. Playing in general 1500+ games (to be precise, average opponent ELO 1605). And I am a little bit lazier than Drake, so I threw wins and ties together.

My data based on 243 games:

Started 121 games
Won or tied 83 games (69%) [D(rake) 70%]
Won with longest 64 games (77% of won games) [D 76%]
Lost 38 games (31%) [D 30%]
Lost with longest 12 games (32% of lost games) [D 31%]

Not-started 122 games
Won or tied 72 games (59%) [D 49%]
Won with longest 50 games (69% of won games) [D 75%]
Lost 50 games (41%) [D 51%]
Lost with longest 10 games (20% of lost games) [D 31%]

So, preliminary conclusions for me:
- surprisingly similar results for the games we both started
- significant performance difference in the not-started games; my data seem roughly in line with Angels assessment of the advantage for the starting player, whereas Drakes data suggest much more difference between starting or not-starting
- interestingly Drake seems more capable of attaining longest in his not-starting games, but seems less able to convert them into a win; based on the starter games our play style should be roughly the same, maybe the 1400+ part of Drakes stats includes more ticket drawing opponents which is more advantageous for them in started games? Or ticket drawing grew more popular in 1 year time?

And yes, Bülent, also a lot of Excel involved over here....
      
UEG xbomanx
Senior Member
T2R Single Player World Championship 2010 Winner

User Pages
Posts: 339
Registered:
May 2008
Re:On the importance of starting Sun, 26 February 2012 23:02
Thanks Patrick very meaningful. Nice work guys.

Here is my opinion:

we assume that both players have about the same ELO-score wins in the ratio 60-40% of the players whose the start has.

same number of points = A vs. B = 20 (60% -40%)

we assume that a player 50 points more than his opponent, he wins in the 65-35% comparatively to the start.

so = A vs. B = 30 (65% -35%) = +50 differential

now assume that a player 100 points more than his opponent, he wins in the 70-30% comparatively to the start.

so = A vs. B = 40 (70% -30%) = +100 differential

I could now carry on an ongoing basis. but how it is when a player has 50 of the start points ELO has less than his opponent?

I think 50-50%

then A-50 vs. B = 0 (50% -50%)

if a player has 100 ELO points less than his opponent, he loses in the ratio 45-55%, despite start.

then A-100 vs. B = -10 (45% -55%)

difference of 150 points in a 40-60% would be comparatively.

then A-150 vs. B = -20 (40% -60%)

also I could now carry on an ongoing basis.

Result: is the skills of one player better, the more he wins. Since it does not play a big role whether you start or not. Above all, it is also on the abilities of the opponent. So i can agree with Angel and the statistic of Drake and Patrick, cause they say the same. Whatever have fun all and not to much thinking about starting or not.

These results are based on over 3000 games only for 2 players.

Many greetings
xbomanx




      
ATN Drake
Senior Member

Posts: 819
Registered:
March 2006
Re:On the importance of starting Sun, 26 February 2012 23:35
DD Patterson wrote on Sun, 26 February 2012 06:29

So, preliminary conclusions for me:
- surprisingly similar results for the games we both started
- significant performance difference in the not-started games; my data seem roughly in line with Angels assessment of the advantage for the starting player, whereas Drakes data suggest much more difference between starting or not-starting
- interestingly Drake seems more capable of attaining longest in his not-starting games, but seems less able to convert them into a win; based on the starter games our play style should be roughly the same,


The conclusion I draw is that You and Angel are better players than me in 2P USA. And I think it starts to show when you are the non-starter. Going first you have more room for error and still able to pull out a win. When you go second, any mistake or just not playing as well hurts you, and it starts to show in the long run.

      
    
Previous Topic:2ND story again
Next Topic:Second Account - Accusations
Goto Forum: