Goal - The Dream Begins

The story of a young mexican player, growing to a star. Created in collaboration with the FIFA, it includes real teams and real player.

En Français:
Goal ! : Naissance d’un prodige

Goal 2 - Living the Dream

Once you are hooked by the first episode, get the 2nd episod of the trilogy (The third episod is currently being produced!)

En Français:
Goal 2 : la consecration

I have received some very interesting information on how PsicoTSI (the skills sub levels tool) works in details and I would like to share it with you. It explains the exact formulas used by Re4Ver when creating it a few years earlier.

A word of warning though, this article is mostly for people understanding a minimum of programming…

Basically, the source code is fully available in the earliest versions of PsicoTSI, a bit of unpacking of the xpi, first and then jar files included in the xpi with your favorite zip tool, and you have the whole source code in front of you.

PSICO TSI main component is a function approximating the TSI calculation, and here is in a few words the PSICO TSI algorithm:

1) open a player page with visible characteristics
2) collect information from the player page (CHPP approval needed!)
3) detect the main skill of the player
4) iterate between sub levels of the main skill
5) estimate the TSI for each sub-level
6) if the closest TSI is found, return the sublevel

The TSI formula is based on neural networks, as explained in this previous blog post.

Applications of this formulas can of course be of extreme help to everyone, and obviously, making it officially open source would provide a great improvement to mainly CHPP applications out there.

For reference, the PSICO TSI javascript code of the TSI estimation part:


function sim(pinput,mainSkill) {
// PlayMaking
if (mainSkill == "PM") {
var meanp = new Array (0,6.37560321715818,6.8295799821269,9.18572832886573,3.82265415549598,3.80478105451296,3.97028596961573,3.7316800714924);
var stdp = new Array (1,1.27931620538792,2.068121874063,2.64814773627783,1.25944526479215,1.23643585187472,1.41074883966057,1.200261576732);
var meant = 3.44951885218574;
var stdt = 0.69620243901134;
var IW = new Array (new Array (-0.978045139728464,0.00301687457750677,-0.048841998968147,-0.404735131313642,-0.0362317679661696,0.0379976393063689,0.010290023976103,-0.0147760361492318),
new Array (0.177344012094537,-0.0314455259871624,-0.013402962769724,-0.0892159635787965,0.149684600060107,0.0129173095791169,-0.0627676826414977,-0.0140624018719683),
new Array (1.36521889783453,-0.00637939702956546,0.0153705440757856,0.556447703939989,-0.0241943951441762,-0.0493823748265197,-0.21427629169483,-0.00101793041393101),
new Array (0.196053426793117,-0.02079754895227,0.00105220616818506,-0.0444692661662335,-0.125304863742313,0.000952177286031863,0.0555651518906449,-0.0497925245034302),
new Array (0.692342189504846,-0.651215170938504,-0.0773280848125734,-0.0299899107255201,0.0102778174399698,-0.00729484262085779,0.019086283284809,-0.0704572408040264),
new Array (1.58818822767738,-0.00884634038142363,0.0255779210123106,0.487506584936556,0.0814314344175245,-0.251017981688899,0.0901880561018751,-0.0166380710765307),
new Array (-0.349984574105507,-0.00620252485894222,0.0100312003087919,0.23260892557538,0.0260982230419322,0.111026396362525,0.037817626462384,-0.319122796193367));
var LW = new Array(1.45005860984466,-5.08419804987079,-3.09594267675791,-2.19350882485451,-3.55681002955351,0.212673777310428,-2.17305546140852,0.669079546193561);
var phidden = new Array (1,0,0,0,0,0,0,0);
for (var i=1; i<8; i++) {
for (var j=0; j<8; j++) {
phidden[i] = phidden[i]+((pinput[j]-meanp[j])/stdp[j])*IW[i-1][j];
};
phidden[i]=tanh(phidden[i]);
};
var poutput = 0;
for (var k=0; k<8; k++) {
poutput = poutput + phidden[k]*LW[k];
};
return Math.pow(10,poutput * stdt + meant);
};
// Winger
if (mainSkill == "WG") {
var meanp = new Array (0,6.11559888579387,6.38544568245125,4.19986072423398,9.07820334261824,3.78899721448468,4.09261838440111,3.94707520891365);
var stdp = new Array (1,1.35219800030418,1.87676157709407,1.60495529959243,1.92480738248487,1.23672495343354,1.35904038569407,1.42286644873124);
var meant = 3.34828115361494;
var stdt = 0.50248598993484;
var IW = new Array (new Array (-0.660772073958308,-0.01833440065284,-0.0878004061319683,0.00014551774875378,-0.09581305124871,-0.018902774261109,-0.0267581613703629,-0.0533105102621847),
new Array (0.512421132362667,0.0382337124466166,0.0773884058542099,-0.0478971000074065,-0.227090911654096,0.0749380149540642,-0.0192496230658901,0.591939574122913),
new Array (-1.37612347434054,-0.0172255940860493,0.0372075456228269,0.732664914012482,-0.636332206449557,0.0660828680402308,-0.13569710277901,-0.0969683305370428),
new Array (-0.767604389511557,0.0164740614163994,0.0996416926112712,0.0574768132107605,0.406081043832447,-0.0338799984024221,-0.755504987651843,0.0404297482763663),
new Array (0.51444631220829,0.0227910340702376,-0.194434226880239,0.0157033323422596,0.145696031716467,0.0280447811851727,0.0997626514667552,0.0600548994144656));
var LW = new Array(-4.87066368568217,-8.12793906828807,-0.56007809906691,0.451613517207488,0.354350670244603,2.07516604461788);
var phidden = new Array (1,0,0,0,0,0);
for (var i=1; i<6; i++) {
for (var j=0; j<8; j++) {
phidden[i] = phidden[i]+((pinput[j]-meanp[j])/stdp[j])*IW[i-1][j];
};
phidden[i]=tanh(phidden[i]);
};
var poutput = 0;
for (var k=0; k<6; k++) {
poutput = poutput + phidden[k]*LW[k];
};
return Math.pow(10,poutput * stdt + meant);
};
if (mainSkill == "SC") {
var meanp = new Array (0,6.21562558619396,5.79731757643969,3.79764584505721,3.78432751828925,9.20751266179118,4.23921403113862,3.88655974488839);
var stdp = new Array (1,1.45803238481507,1.84919443347091,1.29668412881596,1.21518918313263,1.71476790521195,1.54547609249067,1.29572069955967);
var meant = 3.45478105603469;
var stdt = 0.44019807215147;
var IW = new Array (new Array (1.40301200769146,-0.0193144940635949,0.0212311747405729,-0.175540588154821,-0.0221115046234967,0.214900394271821,0.0193996813745877,-0.0481966024580163),
new Array (-1.30320677336796,-0.00823005974306588,-0.0639798654429837,-0.0548890263553145,-0.000909477091528763,-0.439094120327123,0.616393481014018,-0.00643510347147235),
new Array (0.622274133245999,0.0155796819413849,0.231654894650702,-0.0891031925337777,-0.157262349055432,-0.589802519029839,-0.143798421154869,-0.187230559566363),
new Array (-0.634112378069713,-0.000944335329076177,-0.0544022297082993,0.0193751289262962,-0.0100586795623683,-0.106747528361095,9.52666848716332e-005,-0.035002277549141),
new Array (-0.760416443748201,-0.00710422254483257,-0.0438615652834481,0.102521784762147,-0.0281872796781719,-0.0687360185898756,0.0163874310887414,-0.228778191132918),
new Array (-2.04885929453845,-0.320961567561186,0.000944441355045404,-0.0305253294471756,-0.00788624645149755,0.0126358849478757,0.00808561844967016,-0.00871625102475451),
new Array (0.283890100480124,0.0136008146777745,-0.362135507458957,-0.00175774533580776,0.0181290187333429,0.220512760178359,0.044816292969784,0.032774317851683));
var LW = new Array(-7.7074616828558,-4.81464556813888,0.640026643980887,0.221373217042846,-17.1398121382442,2.59419730774328,-4.26616397837132,1.10880653486695);
var phidden = new Array (1,0,0,0,0,0,0,0);
for (var i=1; i<8; i++) {
for (var j=0; j<8; j++) {
phidden[i] = phidden[i]+((pinput[j]-meanp[j])/stdp[j])*IW[i-1][j];
};
phidden[i]=tanh(phidden[i]);
};
var poutput = 0;
for (var k=0; k<8; k++) {
poutput = poutput + phidden[k]*LW[k];
};
return Math.pow(10,poutput * stdt + meant);
};
if (mainSkill == "PS") {
var meanp = new Array (0,5.9037558685446,6.04694835680751,4.12323943661972,4.09976525821596,4.21713615023474,7.21807511737088,4.38615023474178);
var stdp = new Array (1,1.52228260105579,1.95290122400408,1.29144037033675,1.29438644360604,1.49333417169661,0.961187627213399,1.67817798097032);
var meant = 3.06630240883934;
var stdt = 0.42308172264001;
var IW = new Array (new Array (-11.8900701342496,-0.024750835138891,-0.965752396040145,-2.09746089518353,-0.321967528454612,-2.38676435871213,-5.1471801821007,7.79174276667143),
new Array (0.529380858245977,0.0291519802983588,0.0369583467350457,0.032424613698893,0.0329191863310305,0.0396041328798249,0.0930316090276224,0.0356752670740693),
new Array (2.33113350771975,3.74579049731695,7.11876326800507,-7.91110226838671,3.6497359930875,-3.87258670301254,-3.046829371455,5.84457458310682));
var LW = new Array(-3.83433618763949,0.297797052746795,8.55140826231407,-0.0817402547000788);
var phidden = new Array (1,0,0,0);
for (var i=1; i<4; i++) {
for (var j=0; j<8; j++) {
phidden[i] = phidden[i]+((pinput[j]-meanp[j])/stdp[j])*IW[i-1][j];
};
phidden[i]=tanh(phidden[i]);
};
var poutput = 0;
for (var k=0; k<4; k++) {
poutput = poutput + phidden[k]*LW[k];
};
return Math.pow(10,poutput * stdt + meant);
};
if (mainSkill == "DF") {
var meanp = new Array (0,6.71400113830393,5.55150825270347,3.49132043255549,3.53173022196927,3.54083665338645,3.75028457598179,8.03890153670999);
var stdp = new Array (1,1.48828675945146,1.84177487301679,1.14089446002011,1.1547523721879,1.09262970647274,1.27111995210437,1.91951279588256);
var meant = 3.07079169070856;
var stdt = 0.54669631751747;
var IW = new Array (new Array (0.6857172999329,0.144851901928772,0.00732053581209846,0.142004608852104,0.322390526898797,0.541747526748434,0.00470511682361453,-0.352618377429687),
new Array (0.620509248945402,-0.0126640050506412,-0.0313294540757459,0.0726382032031436,-0.0658046623810817,-0.0531578337102174,0.14309876313257,-0.262817885221702),
new Array (0.361958560232547,0.0191873845346643,0.0940541699357571,-0.042618354807883,0.123768814062805,0.178214776227766,0.0870625029013581,0.545801377669313),
new Array (0.807238317377399,0.050804039114291,0.0338844355782919,0.246852103250382,-0.0694942575939604,-0.10098194009457,-0.354751810513075,-0.124777986713548),
new Array (-0.341621423026568,0.000309270803772757,0.00956076398728777,0.175139999720122,-0.0332175436812512,-0.0333469154384476,0.0646493157528646,-0.0239336387359904),
new Array (2.76039548716256,-0.00904355194026333,1.64943269580411,-0.0338036046909946,-0.198593977982261,-0.109517200651956,-0.167444994960601,-0.230909331468331),
new Array (1.58875346153719,0.293418924006587,0.0211646546315773,0.0824895902178759,0.0040093750686147,0.0394329201279045,0.020570514119798,-0.0305463551011085));
var LW = new Array(0.140377452993393,-0.339571400603176,-2.30959398112466,0.985381658295128,-0.767870805617256,2.35039694690841,0.215556071649187,2.22390988552588);
var phidden = new Array (1,0,0,0,0,0,0,0);
for (var i=1; i<8; i++) {
for (var j=0; j<8; j++) {
phidden[i] = phidden[i]+((pinput[j]-meanp[j])/stdp[j])*IW[i-1][j];
};
phidden[i]=tanh(phidden[i]);
};
var poutput = 0;
for (var k=0; k<8; k++) {
poutput = poutput + phidden[k]*LW[k];
};
return Math.pow(10,poutput * stdt + meant);
};
}


Some vintage pictures from my collection:

Luxembourg with 57 users

The very first teams to appear in the Nationaldivisioun.

I really wish to keep contact with all those nice people i got to learn in hattrick.
Good luck guys.

What proportion of the luxembourgish players are now playing abroad? In which country are they now playing?

As you will see not only the players from our national teams have found a home outside of the grand duchy.
Find out more in this study on not less than 8800 players!

Firstly, the study has been done on all non bot teams and the 8800 players found are roughly estimated to be at least 90% of the luxembourgish players playing in active teams.

Where are they playing?

63% are playing in Luxembourg, but that also means 37% are playing in another country

Do we have players in Mongol Uls?

No, no luxembourgish player made it to there. However we have one player in Cabo Verde, Iran, Azerbaidjan, United Arab Emirates, Kyrgyzstan… If you know the player ids of the players, please post them here!

What are the top nations hosting our local players?

Nederland 437 13,31%
België 354 10,78%
España 253 7,71%
Italia 242 7,37%
Deutschland 197 6,00%
Schweiz 165 5,03%
Portugal 132 4,02%
France 126 3,84%
Polska 112 3,41%
Danmark 104 3,17%

How can we explain the rankings?
Personally, I see 3 main reasons which in order of importance would be:
1) The size of a country. Large countries will buy more players, and therefore their share will be higher
2) Transfer region of the country. On the transfer list, the players matching your criteria coming from the same region (Western Europe) will be the first to be seen, and therefore increase the chance of them being bought
3) Interest in the country. People from our neighbor countries who know a bit in real life or have contacts to the country are most likely to follow the local market. The same trend is of course seen with the interest in our national teams, it is for the same reason that Seiei, alfredbester, Dakar became national coaches

Why is Denmark so high in the ranking?

A single team is boosting the rank of Denmark to those heights…
Of course, it is the one of our ex-U20 coach Ostekiks.

Conclusion

It was a fun study to make, and I hope this outside of the box presentation of the transfer market was interesting for you!

Scottish manager Talkingcat said once about AoA that “if your opponent can’t do anything about it, you should use it“, of course this implied that on the other hand the tactic could be countered effectively.

HT-Tjecken reinforced that in 7831771.37: “the AoA tactic’s “success” is more because a lot of people fail tactically (imho) when facing an all out attack team. (…) The all out attack tactic is an extreme one, but it’s not superior. In fact it’s far from superior. And as long as it’s not superior, why should we limit its existence?”.

This seems to imply that we can expect AoA to be there for a long time, and we’d better learn to counter it effectively.

The main weakness of AoA teams is the defense, which will allow their opponent to score nearly all chances if you have a tad of attack.

To counter effectively AoA, the main principles are:
1 - high midfield level, even if you cant get the possession it will help you to have some chances and reduce your opponent’s chances.
2 - build up good defense ratings on the fronts used by the AoA team to reduce the chance that they score their chances
3 - attack on all fronts without necessarily having very high ratings, so that you have good chances to score your own chances.

Type of player orders that can be useful:
- forwards toward the wing: they will allow you to attack on all sides while you will be able to score a lot of the chances
- defensive forwards: they take advantage of the low attack requirements and boost your midfield level
- good WTM: which will help have both a bit of wing attack and boost your midfield

Some additional tips:
- avoid AIM or AOW team orders as they would reduce your defense levels and are unneeded for your attack to be effective
- have good defenders which will help, not to block the strong attacks, but so that you can use some of the other defensive spots to build up your attack or midfield levels

What about CA then?
If you leave midfield completely to your opponent, there is a strong chance that despite very high defense ratings, your AoA opponent will score some goals, and then despite a possibly high CA level, it is lady luck who will be able to save your game.
CA is not the best option to counter AoA teams, except for very few extreme teams which have optimized their teams for great defense, CA and attack levels.

This article was based on contribution from DurdenTyler, and conference posts in particular from HT-Tjecken and Talkingcat, thanks and credits to them !

In that web site you’ll discover a lot of ideas that have been developed by hattrick managers in the whole worls, you’ll be able to comment on them and rank them. Although the web site is not run by the HTs, its structure and collaborative approach make it a very efficient lobby in the hattrick community. Add your voice there will help improving hattrick in the way you would like to see it!

The ranking of the highest salaries are available in alltid.org. Therefore I wont put the whole list I used but only some of the movers.

The below data was computed on the 24th of January at the effective middle of the season (between the 7th and 8th game).

Wage increasing!
F.C. Koeppchen Wor 561220, +387712
Diekrecher Ieselen 462676, +305372
FC Fismoluni 356988, +231652
Heisdorf 363272 , +226732
Ha´-Wa´ 299152, +86408
Belair Allstars 517040, +84320
Power Warriors 213892, +67604
FightClub Football 568760, +61372
LIMPERSBERG FC 454524, +59236
FC Feber 241920, +58652

Reducing bills!
Jeunesse Champion 385908, -114044
Sandweiler 449356, -94440
DreamTeam Esch 165372, -69244
Dynamo Weimershof 106536, -12040
Atletico Noodles 264808, -7924
Dna Crew 201732, -7484
Edinburgh DUST 320780, -4652
Luxembourg Soccer 400516, -2604
Team Ariete 143376, -1440
sandweiler ovals 323488, -680

Nationaldivisioun changes:
Spora 715804 , +860
FightClub Football 568760, +61372
Jeunesse renerts 566380, +51520
Belair Allstars 517040, +84320
Gainesvillerockcity 502724, +3520
fc formatdepoche 497632, +30604
Luxembourg Soccer 400516, -2604
Jeunesse Champion 385908, -114044

This article will describe what it takes to build an efficient All Out Attack (AO or AOA) team so that, may be, you’ll become as succesfull!

How to build the AO team:
- Buy a lot of midfielders, as you will use 7 of them to build up a maximum midfield strength
- Play with 2 offensive central defenders (use midfielders in those spots), they will provide you with the midfield boost you are expecting from such a tactic. The last defender will most probably be played as offensive wingback as he will not be able to close all the defensive gaps you will have.
- Can use offensive wingers if you know that your midfield will be dominant enough, however you take a risk then if you are not so sure about your midfield dominance.
- You will be able to play both Attack in the Middle (AIM) or, if you have wingers, Attack on Wings (AOW) if your opponents defensive lineup presents some obvious holes!

Advantages:
- Very strong midfield and attack allows you to score a lot of goals, and as you are ensuring very high possession levels, your opponent will not have many chances (virtual defense)
- Optimized for Playmaking trainers, who are able to spare buying substitutes to train as you will have players playing offensive defenders in the league game and be midfielder in the friendly game

When to use AO:
- Basically, if your opponent does not have enough defense and midfield to stop you then you can use this tactic. It will help give a good goal average if you are playing against a weaker team.
- It can also be used to surprise stronger teams which will underestimate you.

Conclusion:
This tactic showed a lot of efficiency however it must be said that the teams who are relying exclusively on the AO tactic are enjoying that success mostly in lower level leagues (also some teams have also been successfull with it at higher levels).
Many managers find this tactic boring (citing HT-Tjecken) as it is very far from reality and because so many teams are relying completely on it instead of relying more balancing various tactics.

Thanks to DurdenTyler for his input on this article, I hope for a long and successful collaboration writing articles with him!

Don’t miss the next article on countering all out attack!.

This has been confirmed in HT-Tjecken’s statement in conference post 7963000.878.

A whole list of Regainer players and some background information is available on hattrick’s wiki.