Forse qualcuno si ricorderà che lo scorso giugno avevo abbozzato una ricerca su quali Traders italiani avevano realizzato i migliori guadagni nei primi 5 mesi del 2010. Bene, terminato il 2010 e trovato un po’ di tempo, ho deciso di ampliare l’analisi a tutto lo scorso anno solare, in modo da avere un campione più ampio su cui fare delle considerazioni.
Scopo della ricerca è:
1) draw up a ranking of users who have obtained the best gains in 2010
2) to analyze in detail how the best traders are moving, going to see how to distribute purchases and sales in several weeks of the season and what kind of price they are working . In this second part of the research, I had to keep in mind that the calendar year 2010 does not match an integer number of seasons hattrickiane: start with the first 41 weeks of the season, continues with the 42 and 43 and the remaining two are of 5 weeks Season 44. To get an accurate picture of seasonal movements so I have excluded from this part of the 44 weeks.
Given the sensitivity of the issue, I put subito le mani avanti con un po’ di doverose precisazioni metodologiche.
Punto PRIMO : questa non è “LA” classifica assoluta, ma UNA delle classifiche possibili . Ci sono molti modi infatti per valutare quali siano i guadagni dei manager, da quelli meno raffinati a quelli che spaccano il capello in quattro, tuttavia anche questi ultimi presentano delle imperfezioni, per tutta una serie di fattori che vado subito ad elencare:
A) per avere un dato più pulito, si potrebbero considerare solo i giocatori effettivamente comprati e venduti nell’arco di tempo in esame, escludendo quelli comprati nel 2010, ma non ancora venduti, e quelli venduti nel 2010, ma comprati precedentemente. E qui si trova subito un grosso ostacolo: non sono pochi i giocatori che per un motivo o l’altro sono stati licenziati: mettiamo che io venda a 2kk un 34enne, l’utente successivo poi lo licenzia, io i soldi li ho ricevuti, ma dalla storia trasferimenti risulta la dicitura “licenziato” che impedisce di valutare correttamente la compravendita del giocatore (se sono presenti più giocatori licenziati non è possibile sapere con certezza a quale acquisto corrisponde quale vendita, in quanto compare solo la dicitura “licenziato”). E’ una distorsione non da poco. Una soluzione semplice consiste nel considerare l’insieme di tutti i giocatori comprati e venduti nel 2010, presumendo, roughly, that bought in 2010 and not yet sold will go some way to compensate with sales in 2010 but bought before.
b) salaries. Take a player X, bought and sold 1kk 2kk after 15 weeks if his salary is 10k net weekly gain of wages is (2000-1000-10 * 15) = 850K, but if his salary is 100k weekly gain is actually a net loss (2000-1000-100 * 15) =- 500k. Here are several issues: salaries, as we know, vary for each player's birthday and to calculate the net should be going to see week by week on Alltid salary effective debris, an immense task but would be the usual limit of the players fired: what was their salary? We do not know and we can not know, then a distortion, however, would remain. For that purpose I propose and for the time at my disposal I go for the easiest solution, that is considered gross salary, which implicitly overstates the earnings of managers that buy and sell players with high salary.
C) fees. As you know the fees vary according to length of the stay of the player in their team, lower day and higher taxes. Here, too, has the usual problem of redundant players that we can not calculate with certainty the number of days of stay. A further distorting element is given by income as "former owner" should go see player for player if it was then sold in the time and since, however, creating an asymmetry between those that are sold and those that are instead kept in the team the following year. Here too, for simplicity I considered before tax.
I then considered the set of trades made in 2010, without ruling out the players bought and sold in 2010 and sold in 2010 but bought before and, as stated, inclusive of salaries and taxes.
SECOND point: I mentioned in the introduction of "an analysis of what Italians Traders had made the best gains, "attention to the words I did not say the" best traders ", since the ranking that follows is not caused by a competition among traders, but a survey from 1/1/2010 to 31 / 12/2010. Only if users were aware of this survey would be pointless to assess the results as a classification type race since then was put in place strategies to maximize profits over the period, while this is a "picture" with dates arbitrary users who were not trying to give the maximum in the period under review.
For the detection mode also performed there may be (and indeed there are) the distortions mentioned above: if a user buys three players in each other's 3kk 30/12/2010 maybe superaffaroni 4kk to sell to one after a few weeks, the data will see them simply as a total loss and gain the user, the difference between all the players of those bought and sold, will be reduced by 9kk. The reverse situation if you had sold 2/1/2010 several players bought before. As mentioned above I have assumed a flat rate that these situations are balanced.
Another element is that users who leave or re-casting (Daveheart, Kaleir?) So that by selling all or almost Players create the park is a big difference between sales and bought players. It is not clearly gains from trading and situations are identified in the table just by the strong asymmetry between the number of players bought and sold many players. Public
the data processed so far:
analyzed a total of 19,037 transfers for a total of almost € 18 billion
Scopo della ricerca è:
1) draw up a ranking of users who have obtained the best gains in 2010
2) to analyze in detail how the best traders are moving, going to see how to distribute purchases and sales in several weeks of the season and what kind of price they are working . In this second part of the research, I had to keep in mind that the calendar year 2010 does not match an integer number of seasons hattrickiane: start with the first 41 weeks of the season, continues with the 42 and 43 and the remaining two are of 5 weeks Season 44. To get an accurate picture of seasonal movements so I have excluded from this part of the 44 weeks.
Given the sensitivity of the issue, I put subito le mani avanti con un po’ di doverose precisazioni metodologiche.
Punto PRIMO : questa non è “LA” classifica assoluta, ma UNA delle classifiche possibili . Ci sono molti modi infatti per valutare quali siano i guadagni dei manager, da quelli meno raffinati a quelli che spaccano il capello in quattro, tuttavia anche questi ultimi presentano delle imperfezioni, per tutta una serie di fattori che vado subito ad elencare:
A) per avere un dato più pulito, si potrebbero considerare solo i giocatori effettivamente comprati e venduti nell’arco di tempo in esame, escludendo quelli comprati nel 2010, ma non ancora venduti, e quelli venduti nel 2010, ma comprati precedentemente. E qui si trova subito un grosso ostacolo: non sono pochi i giocatori che per un motivo o l’altro sono stati licenziati: mettiamo che io venda a 2kk un 34enne, l’utente successivo poi lo licenzia, io i soldi li ho ricevuti, ma dalla storia trasferimenti risulta la dicitura “licenziato” che impedisce di valutare correttamente la compravendita del giocatore (se sono presenti più giocatori licenziati non è possibile sapere con certezza a quale acquisto corrisponde quale vendita, in quanto compare solo la dicitura “licenziato”). E’ una distorsione non da poco. Una soluzione semplice consiste nel considerare l’insieme di tutti i giocatori comprati e venduti nel 2010, presumendo, roughly, that bought in 2010 and not yet sold will go some way to compensate with sales in 2010 but bought before.
b) salaries. Take a player X, bought and sold 1kk 2kk after 15 weeks if his salary is 10k net weekly gain of wages is (2000-1000-10 * 15) = 850K, but if his salary is 100k weekly gain is actually a net loss (2000-1000-100 * 15) =- 500k. Here are several issues: salaries, as we know, vary for each player's birthday and to calculate the net should be going to see week by week on Alltid salary effective debris, an immense task but would be the usual limit of the players fired: what was their salary? We do not know and we can not know, then a distortion, however, would remain. For that purpose I propose and for the time at my disposal I go for the easiest solution, that is considered gross salary, which implicitly overstates the earnings of managers that buy and sell players with high salary.
C) fees. As you know the fees vary according to length of the stay of the player in their team, lower day and higher taxes. Here, too, has the usual problem of redundant players that we can not calculate with certainty the number of days of stay. A further distorting element is given by income as "former owner" should go see player for player if it was then sold in the time and since, however, creating an asymmetry between those that are sold and those that are instead kept in the team the following year. Here too, for simplicity I considered before tax.
I then considered the set of trades made in 2010, without ruling out the players bought and sold in 2010 and sold in 2010 but bought before and, as stated, inclusive of salaries and taxes.
SECOND point: I mentioned in the introduction of "an analysis of what Italians Traders had made the best gains, "attention to the words I did not say the" best traders ", since the ranking that follows is not caused by a competition among traders, but a survey from 1/1/2010 to 31 / 12/2010. Only if users were aware of this survey would be pointless to assess the results as a classification type race since then was put in place strategies to maximize profits over the period, while this is a "picture" with dates arbitrary users who were not trying to give the maximum in the period under review.
For the detection mode also performed there may be (and indeed there are) the distortions mentioned above: if a user buys three players in each other's 3kk 30/12/2010 maybe superaffaroni 4kk to sell to one after a few weeks, the data will see them simply as a total loss and gain the user, the difference between all the players of those bought and sold, will be reduced by 9kk. The reverse situation if you had sold 2/1/2010 several players bought before. As mentioned above I have assumed a flat rate that these situations are balanced.
Another element is that users who leave or re-casting (Daveheart, Kaleir?) So that by selling all or almost Players create the park is a big difference between sales and bought players. It is not clearly gains from trading and situations are identified in the table just by the strong asymmetry between the number of players bought and sold many players. Public
the data processed so far:
analyzed a total of 19,037 transfers for a total of almost € 18 billion
As mentioned above this is only one classification.
For example it would be possible to do a ranking of
GROSS PROFIT / SALES NUMBER
that returns the average earnings for each player sold.
In that case you would see those who, in a sense, get more with less effort.
Just out of curiosity that ranking would be (excluding rifondanti or abandoning Astrax, Daveheart, AndR80ea, Altares, and Kaleir ing_Laurentio): 1,
PiC74
with € 938,512 2nd € 497,587 with eliot68
3 ° Supergiampi with € 480,875
_HLaW_ 4 ° with 5 °
€ 451,789 with € 451,694 13Giacomo13
6 ° Stefanick
7 ° with € 447,902 with € 410,807 _Peo_
8 ° _Nick with € 394,228
9 ° PinkMoon with € 385,930
10 ° Marteam2004 with € 357,002
course you can also do the ranking of users who have more buying and selling transactions: 1,
tune 860 operations
2 ° regia5 788
Sir_QQ 736 3 ° 4 °
zagortenay 644
5 ° -HM-568 Toto
6 ° eliot68 564
_Peo_ 7 ° 8 ° 538
wren 520
9 ° rusty83 483
10 ° Alerish 463
For example it would be possible to do a ranking of
GROSS PROFIT / SALES NUMBER
that returns the average earnings for each player sold.
In that case you would see those who, in a sense, get more with less effort.
Just out of curiosity that ranking would be (excluding rifondanti or abandoning Astrax, Daveheart, AndR80ea, Altares, and Kaleir ing_Laurentio): 1,
PiC74
with € 938,512 2nd € 497,587 with eliot68
3 ° Supergiampi with € 480,875
_HLaW_ 4 ° with 5 °
€ 451,789 with € 451,694 13Giacomo13
6 ° Stefanick
7 ° with € 447,902 with € 410,807 _Peo_
8 ° _Nick with € 394,228
9 ° PinkMoon with € 385,930
10 ° Marteam2004 with € 357,002
course you can also do the ranking of users who have more buying and selling transactions: 1,
tune 860 operations
2 ° regia5 788
Sir_QQ 736 3 ° 4 °
zagortenay 644
5 ° -HM-568 Toto
6 ° eliot68 564
_Peo_ 7 ° 8 ° 538
wren 520
9 ° rusty83 483
10 ° Alerish 463
Come detto sopra in questa ricerca vi è una seconda parte in cui vado ad analizzare in dettaglio come si muovono i migliori trader, andando a vedere come distribuiscono acquisti e vendite nelle diverse settimane della stagione e le tipologie di prezzo su cui lavorano.
I primi 5 in classifica sono: eliot68, Sir_QQ, _Peo_, PiC74 e Supergiampi, andiamo a vedere come si sono mossi sommando tutti i movimenti effettuati nelle stagioni 41, 42 e 43. In particolare le compravendite sono state sommate settimana per settimana: sommo la 1a settimana della stagione 41 con la 1a della stagione 42 e con la prima della 43, ottenendo tutti i movimenti effettuati dai manager nelle settimane numero "1" delle stagioni, e così per tutte le altre settimane.
I'm going to go up and start from the 5th, Supergiampi , representing the graph of purchases (in red) and sales (in green) in number and value
I primi 5 in classifica sono: eliot68, Sir_QQ, _Peo_, PiC74 e Supergiampi, andiamo a vedere come si sono mossi sommando tutti i movimenti effettuati nelle stagioni 41, 42 e 43. In particolare le compravendite sono state sommate settimana per settimana: sommo la 1a settimana della stagione 41 con la 1a della stagione 42 e con la prima della 43, ottenendo tutti i movimenti effettuati dai manager nelle settimane numero "1" delle stagioni, e così per tutte le altre settimane.
I'm going to go up and start from the 5th, Supergiampi , representing the graph of purchases (in red) and sales (in green) in number and value
In terms of purchases, Supergiampi avoid like the plague the first few weeks of market in 3 seasons, no, I say no purchase in weeks 1 and 2 of the season. True that in 2010 are the weeks that coincided with the Christmas holidays, bridges and spring holidays in August, but I doubt it is coincidence. Timid purchases in the weeks 3 through 6 and the last of the season, but mostly those from 8 to 13.
sales are essentially constant, except the value basso della settimana 1.
Proseguo con PiC74
sales are essentially constant, except the value basso della settimana 1.
Proseguo con PiC74
che presenta variazioni stagionali poco sensibili, a parte il calo degli acquisti a inizio stagione e quello delle vendite a fine stagione. Ecco i grafici relativi a _Peo_
anche per lui vendite più basse della norma nelle settimane di fine stagione (15 e 16) mentre salgono gli acquisti (soprattutto nella 15), che sono invece più bassi nelle prime 3/4 settimane della stagione. Da rilevare anche il valore altissimo di vendite nella settimana 1 delle tre stagioni in esame e anche il fatto che nelle settimane dalla 4 alla 10 la somma degli buy practically coincides with that of the sales. Sir_QQ
presents graphs of sales cycle, but continuous decline throughout the season, thus concentrating sales in the market beginning your season, while purchases are essentially constant. Finally eliot68
which presents as _Peo_, a great exploits of sales in the first week of the season, while the rest of the season purchases and sales values \u200b\u200bare constant and all in all not too dissimilar. Finally an eye to the profiles of players bought and sold: the question is "Traders with better gain work on the same types of players or not?"
I divided the players bought and sold in 2010 price range of € 125,000 each, so the first band is between 0 and € 125,000, the second between 125,000 and 250,000 € and so on ...
Later will use the shorthand notation, and that is € 125,000 = 125k. Let
profiles, starting from the 5th and that is Supergiampi , rapprensentanto left, in red, shopping and right, in green, sales:
Supergiampi seems to work very specialized: you see the red histogram in the left panel that identifies about 48 purchases made in the range between 1.000ke 1.125k. sales appear to the right of distribution, with a peak in the range 1.375k/1.500k however, suggests that an activity of trading skill between 1,100 and 1.400k.
continues with the 4th, PiC74
traders who concentrate their purchases among 1.000ke 1.600k. The selling was well defined peaks in the three bands just below the values \u200b\u200b"round" of 2.000k, 3.000k 2.500ke. Indeed, there are fewer than 12 sales value of € 1,989,520, € 2,489,520 and 10 to 12 others to 2,989,520. Evidently the "89,520" final wears well. Here _Peo_
purchases are distributed with a curved "bell" with the higher frequencies in the area 1.000/1.500k, a distribution not too dissimilar from those sales, which have peaks around the values \u200b\u200byes "round" of 1kk/1.5 kk/2kk/2.5kk, but taken together seem to configure the same shape "bell" of the purchases, but moved to the right Sir_QQ instead has a distribution completely different
purchases have 3 huge peaks in the range between 125k and 500k (well over half of purchases), histograms with gradually decreasing with increasing value of the players. 3 peaks for sales in the area between 375K and 750k, with higher values \u200b\u200bbut also in the bands a bit 'higher. Finally, the trader with the best difference between sales and purchases, eliot68
's trading that has less specialized, the curves show peaks, but forms a substantially bell indicate that goes to trade on different types of players. Summing up the analysis in detail of these five traders, it seems to me that there is no one "way" trading and that excellent results are obtained with different strategies, both in terms of seasonal patterns, which especially that of the type of players on which to trade.
PS. dai un occhio all' INDICE del blog, ci sono parecchi articoli che ti potrebbero interessare.
Andreac (team ID 1730726 in Hattrick)
Andreac (team ID 1730726 in Hattrick)
This opera by Andreac is licensed under a Creative Commons Attribuzione-Non commerciale 3.0 Unported License . Cioé questo lavoro può essere liberamente copiato, distribuito o modificato senza espressa autorizzazione dell'autore, a patto che l'autore sia chiaramente indicato e la pubblicazione non sia a fini commerciali.
0 comments:
Post a Comment