Come valutare un trading system?

x Luonto e Gioric51

Non è mica detto che i risultati minimi siano <0 solo perchè li abbiamo chiamati Loss probabili. POSSONO esserlo ma possono anche essere delle quantità molto piccole ma positive. Dipende dalla media di tutti i profit per trade e dal loro scarto.

Ciao.
 
X Scalpo
mi sa che stiamo facendo un pò di confusione
la media si calcola sui trades fatti (o meglio, simulati dal TS) e i loss vanno col segno meno
Ciao
 
Esempio Tim 1/10/99 - 18/6/01

Questo è un'esempio riferito al post precedente:

Equity profit
xtrade n

0,019 0,019 1
0,118 0,099 2
0,235 0,117 3
0,398 0,163 4
1,095 0,697 5
1,564 0,469 6
2,951 1,387 7
3,098 0,147 8
3,690 0,592 9
4,063 0,373 10
3,784 -0,279 11
3,383 -0,401 12
3,757 0,374 13
5,111 1,354 14
4,914 -0,197 15
5,304 0,390 16
6,289 0,985 17
6,337 0,048 18
6,583 0,246 19
7,876 1,293 20
8,414 0,538 21
6,702 -1,712 22
8,022 1,320 23
8,123 0,101 24
8,092 -0,031 25
8,061 -0,031 26
8,075 0,014 27
7,687 -0,388 28
8,740 1,053 29
9,356 0,616 30
8,755 -0,601 31
9,052 0,297 32
8,877 -0,175 33
8,987 0,110 34
8,921 -0,066 35
10,186 1,265 36
10,360 0,174 37
10,062 -0,298 38
10,256 0,194 39
10,203 -0,053 40
10,446 0,243 41
10,715 0,269 42
11,522 0,807 43
11,710 0,188 44
11,821 0,111 45
11,933 0,112 46
11,617 -0,316 47
11,662 0,045 48
11,834 0,172 49
12,232 0,398 50
12,258 0,026 51
12,008 -0,250 52
11,659 -0,349 53
11,318 -0,341 54
11,474 0,156 55
11,714 0,240 56
11,629 -0,085 57
11,052 -0,577 58
11,281 0,229 59
11,741 0,460 60
11,777 0,036 61
11,502 -0,275 62
11,387 -0,115 63
10,977 -0,410 64
11,207 0,230 65
11,347 0,140 66
11,120 -0,227 67
11,231 0,111 68
11,191 -0,040 69
11,156 -0,035 70
12,567 1,411 71
12,662 0,095 72
12,852 0,190 73
13,005 0,153 74
13,334 0,329 75
13,753 0,419 76
13,853 0,100 77
13,763 -0,090 78
14,325 0,562 79
15,010 0,685 80
15,390 0,380 81
15,620 0,230 82
15,523 -0,097 83
16,183 0,660 84
15,883 -0,300 85
16,001 0,118 86
15,741 -0,260 87
15,911 0,170 88
16,521 0,610 89
16,212 -0,309 90
16,390 0,178 91
16,660 0,270 92
16,659 -0,001 93
16,919 0,260 94
17,054 0,135 95
17,324 0,270 96
17,254 -0,070 97
17,166 -0,088 98
17,349 0,183 99

Analisi della serie storica:

op. fatte 99
op. in utile 66,67% 66
somma dei profit 17,349
max Gain 1,411
max Loss -1,712
media profit per trade 0,18
dev. std 0,47
n 99
gradi di libetà 98
t_0,975 2,08
probabile profit max (Gain) 0,27
probabile profit min (Loss 0,08
prob max/min 3,52
nº max di perdite consecut. 3
nº medio perd. consecutive 2,28
reddito lordo ------
tasse su capital gain ------ commissioni ------
reddito netto ------
capitale iniziale ------ capitale finale ------
rend netto di periodo ------
periodo netto in giorni ------
rendim. med.giornaliero ------
rendim. medio mensile ------ rendim. netto medio annuo ------
velocità dei profit / trade 0,15
 
X Scalpo
OK, però
100 op sono poche per valutare un TS
mi disturba il max loss>del max gain
qual'è il drawdown, il profit factor , il return on account?
 
Gioric51:
Quello che ho riportato è solo un'esempio della metodologia De Lorenzo la quale mi sembra sia sottoposta a critica...
Oltretutto lui, nel suo test usa la distribuzione t_student che fà riferimento alla teoria dei piccoli campioni, cioè n<100.
Più o meno tutti noi siamo consapevoli dell'importanza della numerosità campionaria ma: quanto deve essere grande il campione? 500, 1000, 3000 operazioni? Non sono forse sempre molto poche rispetto al suo universo? E quanto sarebbe grande questo universo?? Credo che questi possano
essere dei punti di riflessione importanti ai quali tenterò di dare una risposta non soggettiva ma che abbia un fondamento nella statistica che io conosco (non è molta è solo l'esame di statistica che ho dato ad Economia e Commercio, non sono di certo uno specialista ma meglio che niente).

Per quanto riguarda i marcatori mancanti da te menzionati, non posso che essere d'accordo con te e ricordo che un TS deve essere testato su diversi punti di vista:

1) Statistico (quello che stiamo facendo)
2) Eco/Finanziario (quello a cui ti riferisci tu)
3) Piscologico (a cui non ha fatto riferimento nessuno)

S'osserva, tra i parametri più diffusi, che il rischio di rovina è un'indicatore strategico.

Innanzitutto rintraccerò le formule di statistica che possano calcolare la numerosità del campione compatibile. Mentre Gioric51, se gli va, potrebbe mettere in chiaro (no Metastock, in chiaro, in modo che tutti possano capire) quelli finanziari.
Nell'attesa di scovare lo psicologo sono attesi suggerimenti.

Grazie in anticipo a tutti voi per il contributo.
 
Allora forse ho capito il motivo di tante incomprensioni.Tu dici:

"Per maggior chiarezza riporto una formula per il calcolo del rapporto G/L:
L_prob= media - dev.std * t/radq(gradi di libertà)
G_prob= media + dev.std * t/radq(gradi di libertà)
un ts è giudicato accettabile se:
G_prob/L_prob >= 3
questo rapporto quindi è effettuato non tra due medie ponderate ma tra due valori probabili.
media= media della serie dei profit per trade
dev. std = scarto quadr. medio "

Nel mio metodo invece opero come segue:
- metto da una parte tutte le operazioni in utile(gain)
- metto da parte tutte le operazioni in perdita (loss)
- mi calcolo di questi due gruppi il rispettivo valore medio e le quantità.
Detto questo mi calcolo l'indicatore che ho accennato sopra.

Nel tuo metodo però L non stà esattamente per loss (!) ma per estremo inferiore del valore probabile di profitto.
Quindi mi rendo conto che siamo su due strade completamente diverse.

A questo punto non mi convince una cosa e di fare un ragionamento per assurdo.
Suppongo di fare 100 operazioni tutte in utile e tutte dell'1%.
La dev.st=0 e quindi L/prob=G/prob e quindi G/L=1.
Come fà ad essere non accettabile un sistema così affidabile?
 
Ma se ho un scarto quadratico pari a zero, dovrei avere una campana che non gli assomiglia affatto, è una riga diritta. Difatti 0*t=0, significa che non ho code!!, neanche l'ombra!! Significa in altri termini che non è una distribuzione di probabilità, bensì una costante. Le distribuzioni di probabilità si applicano a fenomeni casuali, aleatori, variabili, non alle costanti. Alle costanti s'applicano le normali equazioni.
Anche se la statistica risulta comunque matematicamente corretta, nel senso che i valori probabili vengono sempre restituiti in modo corretto. Mentre l'applicazione del rapporto tra questi due valori probabili non è un procedimento statistico in senso stretto ma una ricerca da parte di chi ha proposto il metodo d'un paramentro di confronto, pensando ad una serie variabile. Lo scarto = zero ne annulla il significato, ovvio.
 
Ho precisato che il mio era un ragionamento per assurdo e non troveremo mai nel mondo reale un simile comportamento.
Tuttavia come ben sai i ragionamenti per assurdo servono nel campo matematico a capire relazioni tra variabili.
In particolare non capisco dalla formula che hai postato perchè se la dev.st tende a zero allora G/L tende a 1.

Avere un profitto medio di valore positivo ed una dev.st piccolissima dovrebbe essere un fattore incoraggiante, non trovi?
Perchè con la tua formula mi ritrovo ad avere G/L circa 1?
 
Poichè si ritiene che un TS quando guadagna deve guadagnare molto di più rispetto a quando perde, si è posto come riferimento che quando sussistono dei valori "G/L" di 2-3, bè allora siamo in presenza d'un buon TS. In altri termini, un buon TS non è buono solo se presenta una Equity il più possibile lineare ma deve guadagnare decentemente. Quindi il parametro 2-3 proposto, diventa quasi una questione di gusti, purchè sia > 1 e l'equity non faccia nessun ritorno a zero. Infatti l'equity potrebbe essere inclinata positivamente e presentare sporadici ritorni a zero. Ebbene questi fenomeni sporadici, non dovrebbero proprio esserci.
Da un rapporto pari a 1 e s molto prossimi a zero si direbbe
che quando guadagna lo fa quasi nella stessa misura di quanto perde e questo lo fa in modo costante.
Prova a calcolare l'inclinazione totale della Equity e trasmettimela, insieme alla numerosità del campione e ad i valori max min.

s tende a zero quando le differenze tra i singoli valori e la media della serie tende a zero. Essendo s una misura di dispersione intorno alla media, quando questo è piccolo significa appunto che i tick vinti lo sono costantemente ma la loro ampiezza è anche lei piccola, e ciò dovrebbe anche rallentare la equity.
 
Il fatto che la dispersione sia piccola non implica che la media sia piccola.
Potrei farti molti esempi su questo, e penso che se ci rifletti non potrai che affermare.

Per postare degli esempi, ho necessità di capire perchè nel tuo metodo quando ho dispersione piccola G/L --> 1, mentre quando la dispersione è molto grande G/L --> a valori grandi.

Per maggiore chiarezza ti riporto questi 2 esempi, molto semplificati e giusto per capire:

Esempio 1
Media dei trade 0.80
Deviazione standard 0.05
Quindi G/L -->1

Esempio 2
Media dei trade 0.20
Deviazione standard 0.15
G/L --> a valori grandi (difatti L_prob si avvicina a zero)

E' chiaro a tutti che il TS dell'esempio 1 è più affidabile.

Come mai il G/L del sistema 2 è maggiore di quello 1?
 
es 1 es 2
x 0,8 0,20
s 0,05 0,15
n 500 500
n-1 499 499
t_0975 1,96 1,96
max 0,8101 0,2134
min 0,7899 0,1866
max/min 1,0256 1,1438
2,56% 14,38%


Notare che il rapporto max/min è una percentuale. Quindi
confrontando l'es. 1 col 2 dal punto di vista dell'affidabilità
(che è una qualità che torna utile nei momentanei aumenti del numero di
perdite consecutive nel reale),
sarebbe da preferire quest'ultimo in quanto: quando
guadagna tanto lo fa con il 14,38% in più rispetto
a quando guadagna di meno (o perde). Quindi
questo TS avrebbe le potenzialità di riprendersi prima rispetto al suo concorrente.
Il TS1 invece reagirebbe in modo più blando: con un 2,56%.
Notiamo inoltre che stiamo analizzando degli esempi i cui
guadagni minimi probabili sono entrambe positivi
questo confronto perde quasi di significato se prima non chiarisco
lo scopo del TS che sto progettando.
Voglio dire che se ho un profilo di rischio alto, sarei
più propenso a fare confronti in termini assoluti anzichè in percentuale,
propendendo sicuramente sul TS1 che mi fà guadagnare
di più.
 
Kaufman, che mi pare un pò più autorevole di Di Lorenzo, fa un uso diverso del Student t-Test ; egli infatti dice che un buon sistema deve avere

t= media trades/dev st trades x radice quadr(num trades)>=
della distribuzione t-Student
Col mio Ts viene
t=2,91
mentre dalla tabella t-Student leggo per num trades > 121
t=1,96
per cui avrò 2,91>1,96
Naturalmente questo è solo uno dei pressochè infiniti parametri di cui bisogna tener conto per poter testare un TS, dopodichè occorre accendere un grosso cero alla Madonna e sperare che il cielo ci assista, ne avremo tanto bisogno.

E' un mondo difficile!
Saluti
 
Per validare definitivamente un Ts è necessario predisporre un criterio quantitativo di affidabilità dei risultati ottenuti, rifacendosi a regole di tipo statistico-econometrico.
Fra i vari indicatori, i più significativi in ordine di importanza:

-Annual % gain/loss
-Buy/Hold profit
-Max open trade drw
-Profit/Loss index
-Reward/Risk Index
-Calhoun Profit Ratio = Performance%/peak to through maximum equity drw/n° campioni, deve essere >2 per ritenere il sistema affidabile
-Indice di Sharpe =(Performance% -tasso privo di rischio)/v*performance%, dove v=dev.st. dei rendimenti; il valore atteso è >12

Un aspetto importante da approfondire è la strategia combinata alla probabilità di rovina.
 
Ultima modifica:
x Gioric51:
qual'è il principio statistico che ha guidato Kaufman in questa direzione, o per quale motivo, al di là della sua notorietà, ritieni che questo metodo sia migliore?

Ciao.
 
Il mio dubbio più grande nei riguardi della metodologia De Lorenzo è costituito dall'uso indiscriminato di intervalli di confidenza del 97,5% sempre e comunque. Questo autorizzerebbe chiunque a compiere test con n anche molto piccoli, illudendoli di poter essere confidenti al 97,5%! Cioè con errore del 5% (2,5% per coda).
Difatti la metodologia t_student è vero che è studiata apposta per piccoli campioni (cioè 30-100 casi), ma 100 casi d'un TS fanno parte d'un universo INFINITO!!!
Comunque usando t_student con n>100 s'approssima la distribuzione probabilistica alla normale standardizzata e questo è confortante.
In linea di massima l'errore in statistica si misura con: E=1/radq(n). Quindi se effettuo test con n= 500 E=4,47%. Con questo campione ho un'errore inferiore del 5% e quindi posso potermi permettere per i miei calcoli un'intervallo di confidenza del 97,5%.
Vicerversa, se il mio campione è solo di 100, bè allora l'errore sale a 10%!! Che è una bella differenza.
Visto che questo è un mestiere veramente difficile, si dovrebbe abbondare nella numerosità campionaria almeno fino a 1000. E poi usare nella formula di Di lorenzo il livello di confidenza effettivamente ottenuto. Quello che voglio trasmettervi è che secondo me De Lorenzo mette nella sua formula il livello di confidenza come dato fisso al 97,5%, qualunque sia la numerosità del campione, mentre a me hanno insegnato che
il livello di confidenza è l'analista che lo esprime si, ma con criterio. Quindi perchè non fare coincidere il livello con l'errore derivante dalla numerosità del campione?.
Poi, un'altra cosa: visto che la realtà è al più, piuttosto diversa dal futuro che noi abbiamo creduto di sapere studiare a tavolino, faccio due conti della serva:
se ho 10000 campioni, posso essere confidente al 99%. Se so ad esempio che la realtà mi corregge i miei calcoli all'80%, bè il mio effettivo livello di fiducia potrebbe essere 0,8x0,99=0,792%. Cioè tengo conto d'un ulteriore correttivo empirico per avvicinarmi sempre meglio, alla descrizione della mia funzione probabilistica. E l'uso di questo correttivo è giustificato cari miei, la prossima volta vi dico perchè, ora devo andare.

Bye
 
x Scalpo
Kaufman fa riferimento a lavori di Ben Warwick (Event Trading)
Cmq, Di Lorenzo è veramente ridicolo che scriva un capitolo dal titolo "Valutazione dei TS" e l'unico parametro che prende in considerazione è quello derivato dall'utilizzo della distribuz t.Student quando Kaufman sull'argomento scrive interi libri, non trovi?
 
Here's How to Achieve the Trading Results You Want

Determine whether you are suited to trading

The Key Traits of Successful Traders

Your chances for success as a trader are maximized, if you exhibit most, if not all, of the following characteristics. Take this simple test to see if you have what it takes to win the trading game...

Do you ... Yes No
1. Have a burning desire to trade?
2. Easily take action, once you have made a decision?
3. Have sufficient risk capital (money you don't need to live on)?
4. Possess a strong self discipline?
5. Have a strong sense of self confidence?
6. Have a winning and tested trading strategy?
7. Finish what you start?

Count Yes Answers Chances for Success

1-2 = Poor
3-4 = Fair
5-6 = Good
7 = Excellent!



Renato (mi permetto di chiamarlo così perchè è un mio amico, lo scorso anno per conto del CREDIT AGRICOL INDOSUEZ ( che non è proprio una banchetta...) della quale lui è gestore di GPM, ha ottenuto la bellezza del 38% di rendimento con una borsa che è andata come è andata...
Qest'anno le sue gestioni hanno ottimi rendimenti: Non mi pare che Kaufman possa vantare questi rendimenti!!!!

A voi le conclusioni...


atima
 
questo articolo lo avevo tra gli appunti: buon lavoro

THE TRUTH ABOUT TRADING & TRADING SYSTEMS (and some major reasons most traders end-up losing)

There are a number of reasons why successfully trading commodity futures is so difficult...

Unfortunately, the vast majority of traders lose money ... there are many reasons for this. That's the Bad News. However, the Good News is providing you can get in the small Winner's Circle you too can achieve profitable trading. Then you may reap the rewards with the money lost by the many losing traders flowing to you and richly rewarding you for being a winning trader!

Since there are far less winning traders than losing ones, by being in the winning minority, you will be in a position to receive much greater profits than normally possible! This is especially true what with the great leverage involved with commodities trading, options trading and trading stocks on margin.

Read all about the major reasons so many commodity futures, stock market and options traders lose money (so you can avoid these problems):

After reading this Special Report "The Truth About Trading and Trading Systems" Special Report, you may visit other areas of CTCN's website and our links trading related links for more specific trading knowledge covering all aspects of commodity futures, stocks and options trading for traders & investors.

This Special Report was originally written (by the Editor of CTCN) some years ago. These common trading problems are just as evident (perhaps more so) going into the New Millennium as they were in the early 1990's when this Report was written. The main difference in the report after almost a decade is the fact the prices used as examples may be different. Actually, the price levels themselves make no real difference.

Also, we originally used Daily Bars for time-frame examples, as applied to daily bar charts. Since then daytrading is much more popular with commodity futures, stock market and options traders.

Therefore, we have changed the word "Days" to the term "Bars," as in Bar-Charts. The Time Frames can by identified as inter-day (daily) bars, or intra-day tick-bars, like 1-min., 5-min., 30-min bars (bar-charts), etc. It really makes little, if any difference, as the concepts and theories are basically the same.

Here is The Truth About Trading Trading Systems Special Report:

One reason for losing in the markets is the commodity futures, stocks or options trader is not really sure which time-frame or trend he is trading, or he is not matching his target objective price level to the time frames' expected movement. Perhaps the trader wants to capture a move which he expects to take about 4-bars (or 4-days).

However, the volatility increases so the 4-bar (day) trend is actually over in 2 bars and he does not realize it and stays with the trade 2-bars too long. Thus, he gives up all or most of his profit, because he expected the move to last longer.

The opposite can occur ... this happens when the volatility is low and after just 4 bars he gets tired of waiting for the expected move and exits the trade early, perhaps at a loss or small profit. Suddenly, over the next 2 bars the trend and move he anticipated happens; too late, as he is already out of the trade.

Of course, the 4-bar example above also occurs with traders expecting 2-bar moves which may occur in 1-bar, or vice versa. Also, 6-bar moves which end-up occurring over perhaps 8 or 9 bars, or vice versa, etc., and various intra-day time periods.

Another common occurrence, is the trader not using a specific stop-loss order. Thus, a small loss ends up as a big loss. For example, a trader believes a stop (loss) of $400 is reasonable, based on either technical analysis or on money-management rules. However, perhaps due to discipline problems the trader has, it's not actually used.

Once out $400, he relies on HOPE the market will go back in his direction, and he fails to execute the planned exit point. Frequently, the market fails to move back in a profitable position and the trader is finally forced out of market with perhaps a huge $2,000 loss, instead of the maximum $400 loss anticipated.

Note: More often than not, it seems once the trader who over-stayed the position finally decides to get out, the market frequently reverses the exact day (or next day) he got out! That seems to be an uncanny and almost unwritten law!

The stop-loss order is used, but the stop is not sufficiently precise. More frequently than you can imagine, the stop is hit by just a very small margin. For example, the market may be at 54.60 and a long position stop is placed to sell at 52.50. The market goes down to 52.47 and then reverses back to beyond 54.60 very quickly after stop was barely hit.

Sometimes the stop price of 52.50 may end up being the EXACT low price for that swing . . . very frustrating and upsetting when this occurs! For A Special Report on CTCN's Unique Method for calculating amazingly accurate stop-loss prices - Click-Here for Special Report #2

Why does this happen so often? Because many times the stop-loss price level happens to be a support area based on a trend line, gann angle, old bottom or old top formation, fibonacci numbers, a chart price gap, or just simply an obvious natural stop-loss area, such as a whole or even number.

Thus many other traders use the same logic to place stops at or near the same level. The market gets drawn to that area because that's where orders are sitting that the market (and Floor Traders) wants to get filled. Because of those orders resting in that obvious place, the market price actually moves to that area, almost like magic or magnetic attraction.

Failure to place a stop-loss order with your broker (unless you are always closely following the market using real-time intra-day data, when you are in an open trade) will result in the great likelihood of you losing all or most of your money (eventually) due to one or a couple huge losses caused by the price continuing to drop after going thru your stop-loss price. Sooner or later (probably sooner) it's almost certain to happen, if you don't use and place stop-loss orders to you.

There is an exception for day-traders who are using real-time quotes and watching their real-time quotes and price charts continuously. Sometimes a daytrader may achieve better trading success by using so called mental stops vs. actually placing the stop-loss orders with his broker.

Another reason for failure, is you may be right on a trade, but don't know when to exit the position and take your profits. More often than you would believe, a trader has excellent profits, but ends up giving back all or most of the open equity profits because of not knowing when to get out!

For example, a trader is long at 62.40 and the price moves to 63.90 for a huge open equity profit of say $1,800. He has held the position for a few bars, but after looking at the chart and the powerful up-move, he decides the market should easily go to 64.40 within the next day or two. That way his profit will be $2,500, much more than the current profits of $1,800.

Perhaps the next day the market goes to 64.30 (just slightly under his objective) but ends up closing "weak" because it's "over-bought," and closes for the day at 63.92. The trader is mad about giving up some open profits so hopes it goes back to at least 64.30 again the next day. Unfortunately, some bad news comes out overnight and the market "gaps" down on the next opening and opens at much lower at 63.00.

The trader still hopes for an intra-day rally to get back some of his lost open profits, instead it goes lower all day and the trader finally gives up hope and gets out at a break-even price of 62.40. Because the market was "over-sold," over the next couple bars it eventually recovers back up to the anticipated 64.40 price, but the trader is now out of the market with no profit! This type of scenario is all too common an occurrence. To varying degrees, this happens more often than you would believe!

One solution to this problem is for the trader to take small profits or not use specific targets and place very tight trailing stops just under the market. This is poor practice because you will end up getting stopped out with very small profits most of the time. That will result in your average winning trade being quite small compared to your average losing trade, resulting in poor results.

The best alternative is to use targets scientifically based on the market's volatility. Unfortunately, not many trading systems do that. Ideally, a system should have each and every trade uses a specific and dynamic target price based on the market's actual recent volatility. With a dynamic approach based on volatility and past bar size, the market itself will reveal how far a move should progress, based on actual movement and recent volatility.

Still another reason many traders lose, is because they are using a methodology or trading system which is NOT in actuality fully mechanical, but its trading track-record does not reveal it's not mechanical.

For example, a system's advertisement may claim 60%, 70%, 80%, or perhaps even 90% winning trades. However, these promotional claims are usually based on 20-20 hindsight and subjectivity, and not on real-time actual trades. Perhaps the system says buy/sell when there is divergence between the price and a Stochastics or RSI Study. That divergence is very difficult to recognize in real-time trading, but easy to see with 20-20 hindsight looking at an old chart.

Another popular but subjective approach is to watch for turning points at certain times, also known as time-windows. This approach may say to enter or liquidate the trade after an obvious pivot-low or pivot-high occurs, and providing it's during the projected time-window. It's mostly subjective and easy to do by looking at the past, but hard to do in actual real trading. However, some system developers have in fact used hindsight or subjectivity to arrive at their ridiculous percentage of winning trade claims.

Another popular and over-rated method are Elliott-Wave approaches. Elliott-Wave methods are popular, because there are obviously waves in the markets and the idea of using market-waves to predict market turning-points and also riding these waves, is naturally very appealing to traders.

What I am about to say may upset some proponents of Elliott Waves, but the plain truth is Elliott Waves used by themselves are allegedly of little value in actual trading.

If you want evidence to back-up that statement of their questionable value, just do the following: Show the SAME identical chart to 5 traders (make it a mystery chart - rather than a widely published chart the trader may recall).

Next, ask the 5 traders to specifically define the number of waves they see on the chart. You you will likely end up with 5 different (frequently widely diverse) wave-counts. The chances of even 2 of the traders seeing the same exact elliott-wave-counts are extremely unlikely.

Why is this so? There are in fact "waves" in the markets. However, defining what constitutes a "wave" is near impossible, because a wave is largely a matter of visual interpretation and judgment and is highly subjective. It's difficult for a mostly subjective technical analysis method like Elliott-Wave Counts to be used successfully in trading?

Is there a way to overcome these basically of little, if any value subjective approaches?

Yes! Use a Trading System, which does not rely exclusively on Elliott Waves and other subjective approaches. However, using Elliott Waves as part of a trading plan in conjunction with another time-tested methodology may work for you.

Still another reason traders lose, many follow Time Cycles. Cycles do in fact exist in the markets. For example, Live Cattle may have a reliable long term cycle of 9 to 11-months, low to low. The Stock Market, Wheat or T-Bonds may have a short-term cycle averaging 28 to 30-bars, low-to-low.

The problem is sometimes the cycles may come early or late, or skip a beat entirely. For example, you buy on day number 30 at a price of say 3100, thinking the low is now at hand because the average is 28 to 30-bars and the market closed near its day's high on day-30.

However, because of either fundamental or technical reasons the market's cycle this time will run 35-bars (a common occurrence). During those 5 extra bars, the market goes down sharply to 2900 and below your stop-loss point forcing you out of the trade at a large $1,000.00 loss. Shortly thereafter, the cycle bottoms and the expected move occurs . . . but too late for you because you are out of the market by then!

Alternatively, you buy on day number 30, but you did not realize the low ALREADY happened (4-bars earlier) and at a lower price. You actually ended up buying (without knowing it) 4-bars into the NEW cycle, and at a higher price. Because of that, the market goes up only slightly higher for just 1-bar and then drops sharply because a 12-bar cycle (you perhaps were not aware of or not tracking) is now coming into play, and effecting the 30-bar cycle you are trading. The 12-bar cycle makes the 30-bar drop down and forms a double bottom.

This forces you out of the market because of the sudden loss, stop being hit, or lack of discipline, etc. If your thinking why not trade the 12-bar cycle, forget it, because there's likely a 6-bar cycle effecting the 12-bar cycle, and a 3-bar cycle effecting the 6-bar, etc. Note: Many traders are not aware of the fact there are usually one-half (50%) time cycles within every cycle.

Still another all too common happening is the cycle "skips a beat" and simply disappears for one repetition. The next cyclic repetition works perfectly, but by then you are out of the market with a loss and disgusted and not even following the now working cycle!

Is there a way to solve these problems with cycles? Yes, don't use cycles at all, or perhaps use them in conjunction with other sound methods or technicals.

Still another reason many traders lose, they follow SEASONAL TENDENCIES or subscribe to Seasonal Newsletters, or use Seasonally based Trading Systems. There is no question that Seasonals exist in the markets. This is particularly true in Agricultural where seasonals are very well documented. Seasonals also appear in financials, but not as reliable as agricultural. Seasonals work because of fundamental reasons, frequently tied to the growing season or the weather.

However, it's very hard to make money using seasonal data, regardless of the history of the seasonal tendency. That is because like cycles, sometimes seasonals can be early or late, or worse yet not work at all, also known as a "contra-seasonal move."

A perfect example of a contra-seasonal move, are major bear markets in the grains occurring a couple times during the past several years. According to extensive and well-documented research going back to the 1800's, the grains should move up during late Spring and early Summer. However, at certain times in the 1990's they trended sharply down, when they should have been trending steadily higher according to the seasonals.

Unfortunately, the seasonal experts will be mad about this, but probably the best way to deal with seasonals are to ignore them, especially in markets other than Agricultural markets. Note: Occasionally Seasonal characteristics may be used successfully as a way to enhance or compliment other methodologies. There is no doubt the commercials can tell if the seasonals are early, late or contra-seasonal, but they keep that information to themselves!

Another major problem is you or your system can trade good, but selects the wrong market to trade!

A very common happening. If the market is either far too volatile or on the opposite extreme too dull or flat and sideways, the best system in the world will have great difficulty.

Commonly a system or trader gets "married" to a particular market or market group. For example, perhaps the system is advertised to trade only Coffee or only S&P, etc. The System may do well if that particular market is acting good or trending well or steadily. However, once that market gets either too choppy or flat, that system, no matter how valid the algorithm will very likely lose money or not make money.

What can be done about that problem? Figure out a way to trade only good trending markets. Use software which has a built-in Portfolio Manager and Automatic Trend Ranking Module which selects based on proven scientific methodology, the best markets to trade.

One requirement to do that effectively is to have a trading system which uses the same methodology (patterns) to trade all markets. Still another requirement is the system should be able to select from widely diverse markets. By tracking a number of diverse markets, you can expect about 30% (or more) to be performing or trending well.
 
anche questo mi sembra interessante


Trading Systems Keys

Trading the Turtle system involves many keys. Some food for thought:
Does the 60% drop in NASDAQ stocks mean the bull market has finally run its course? We don?t know, and neither do you. Don't worry about what the markets are going to do, worry about what you are going to do in response to the markets. Turtle trading is not anticipatory. There is no prediction. You cannot predict future prices. No one consistently can predict anything. Prices, not investors, predict the future.
Meticulous risk management strategies are absolutely crucial. Everyone makes money in a bull, but if you don't have a money management plan along the way and an exit plan, you are in trouble when the day of reckoning arrives. Trend following Turtles know when they will get out before they ever get in. They are interested in primarily one variable: price. Forget forecasts, fundamental factors, technological break throughs, the main factor in the sight lines of a Turtle is the price of the market.
Inefficiencies exist in a variety of financial markets around the world that lead to sustained trends. Mechanical trading systems exploit these for big profits. With markets globally in various stages of expansion and equilibrium, a trading strategy must adapt.
Know every day what your portfolio is worth. Calculate what your risks are on any given day for all positions.
Don?t leave stop losses in the market. Mental stops are key.
Controlling risk is not the same thing as avoiding risk. If managing risk is an integral part of your philosophy, when it changes up or down, you simply adjust.
Position liquidations are triggered by significant adverse price action and never at pre-determined objectives. Concentrate on managing the risk. The returns will take care of themselves.
If you are flush with profits you trade in larger size. If you are thinly capitalized you have to cut back. Likewise with risk control. If your stop is very tight, you can put on more contracts at the same level of dollar risk than if your stop is very wide.
Equalized risk means that a fixed dollar amount of risk is allocated to each new position so that for example a corn position will have the same initial dollar risk as a T-Bond position. By trading a system with the same parameters for all markets you protect against when a system works well on historical data, but fails in real time.
You cannot expect to enter a market at the precise moment a bottom is hit, nor will you exit a market at the exact top.
Seek profit opportunities in trending markets, whether those markets are moving up or down.
Obtaining profits from long-term volatility manifested in major price dislocations is the cornerstone of good trading systems.
The objective is not to attempt to buy lows and sell highs, but rather to buy market strength and sell market weakness.
 
x gioric51

prova a darmi tu gli "eseguiti" di 8 operazioni, su 8, intraday in utile (long e short) di seguito alla presenza di osservatori attentissimi:) in una sola giornata

se poi non riesci ad aprire un file....;)

ciiiaaaaaoooooooooo




atima
 
Indietro