{"id":12946,"date":"2026-05-04T14:21:44","date_gmt":"2026-05-04T14:21:44","guid":{"rendered":"https:\/\/bjftradinggroup.com\/?page_id=12946"},"modified":"2026-10-17T22:37:09","modified_gmt":"2026-10-17T22:37:09","slug":"hedge-arbitrage","status":"publish","type":"page","link":"https:\/\/bjftradinggroup.com\/tr\/hedge-arbitrage\/","title":{"rendered":"Hedge Arbitraj\u0131"},"content":{"rendered":"

<\/p>\n

\n
BJF TRADING GROUP \u00b7 TEKN\u0130K REHBER<\/span><\/p>\n

Hedge Arbitraj\u0131nda Sinyal Filtreleme: Difference-to-Open ve H\u0131zl\u0131 Fiyat Ak\u0131\u015f\u0131 Filtresi<\/span><\/h1>\n

\u0130ki brokerl\u0131 bir hedge arbitraj\u0131 kurulumunda tespit edilen fiyat farklar\u0131n\u0131n %60\u201380’i gecikme kaynakl\u0131 yapay sinyallerdir<\/strong> \u2014 ger\u00e7ek LP fiyatlama farklar\u0131 de\u011fildir. Bu sinyallerle i\u015flem yapmak her i\u015flemde para kaybettirir ve brokerlar\u0131n s\u0131n\u0131fland\u0131r\u0131p k\u0131s\u0131tlad\u0131\u011f\u0131 toksik emir ak\u0131\u015f\u0131 \u00fcretir. Bu rehber her iki filtre seviyesini ele al\u0131r: basit difference-to-open e\u015fi\u011fi ve gecikme kaynakl\u0131 kirlenmeyi neredeyse tamamen ortadan kald\u0131ran geli\u015fmi\u015f h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 referans mimarisi.<\/p>\n

2.600<\/strong> kelime
\n7<\/strong> b\u00f6l\u00fcm
\n~12 dk<\/strong> okuma
\nHedef kitle:<\/strong> arbitraj yat\u0131r\u0131mc\u0131lar\u0131, kantitatif geli\u015ftiriciler, broker risk ekipleri<\/div>\n<\/div>\n
\n
%60\u201380<\/span>Filtrelenmemi\u015f sinyallerin gecikme kaynakl\u0131 pay\u0131<\/span><\/div>\n
<%5<\/span>H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresiyle kirlenme oran\u0131<\/span><\/div>\n
30\u2013150ms<\/span>Yapay fark\u0131n tipik kapanma s\u00fcresi<\/span><\/div>\n
3<\/span>Gerekli filtre ko\u015fulu<\/span><\/div>\n<\/div>\n

\u0130ki t\u00fcr hedge arbitraj\u0131 sinyali<\/h2>\n
\n

Hedge arbitraj\u0131 yaz\u0131l\u0131m\u0131 iki broker kotasyonu aras\u0131nda bir fark tespit etti\u011finde, bu fark temelde birbirinden tamamen farkl\u0131 iki nedenden birinden kaynaklanabilir.<\/strong> Birincisi, brokerlar\u0131n likidite sa\u011flay\u0131c\u0131lar\u0131 aras\u0131ndaki ger\u00e7ek bir fiyatlama fark\u0131d\u0131r \u2014 i\u015flem yap\u0131labilir bir f\u0131rsat. \u0130kincisi ise fiyat ak\u0131\u015f\u0131 gecikmesinden kaynaklanan yapay bir farkt\u0131r: brokerlardan biri k\u00fcresel bir piyasa olay\u0131n\u0131 di\u011ferinden daha h\u0131zl\u0131 alm\u0131\u015ft\u0131r ve g\u00f6r\u00fcnen fark yaln\u0131zca daha yava\u015f broker\u0131n kotasyonu hen\u00fcz g\u00fcncellenmedi\u011fi i\u00e7in vard\u0131r. Yaln\u0131zca ham kotasyon kar\u015f\u0131la\u015ft\u0131rmas\u0131na bakarak bu iki durumu ay\u0131rt etmek m\u00fcmk\u00fcn de\u011fildir \u2014 ancak emir ger\u00e7ekle\u015ftirmede tamamen z\u0131t sonu\u00e7lar do\u011fururlar.<\/p>\n<\/div>\n\n\n\n\n\n\n
Sinyal t\u00fcr\u00fc<\/th>\nNeden<\/th>\nKal\u0131c\u0131l\u0131k<\/th>\nEmir ger\u00e7ekle\u015ftirmede<\/th>\n<\/tr>\n<\/thead>\n
Ger\u00e7ek LP fark\u0131<\/td>\nBroker A’n\u0131n LP’si, Broker B’nin LP’sinden farkl\u0131 fiyatlama yap\u0131yor \u2014 piyasa olaylar\u0131n\u0131n zamanlamas\u0131ndan ba\u011f\u0131ms\u0131z<\/td>\n200ms ile 2.000ms aras\u0131<\/td>\nKote edilen fiyatlardan ger\u00e7ekle\u015fir \u2014 beklendi\u011fi gibi net k\u00e2r<\/td>\n<\/tr>\n
Gecikme kaynakl\u0131 yapay fark<\/td>\nBroker A’n\u0131n fiyat ak\u0131\u015f\u0131 piyasa olay\u0131n\u0131 daha h\u0131zl\u0131 ald\u0131. G\u00f6r\u00fcnen fark = fiyat hareketi, LP fark\u0131 de\u011fil<\/td>\n30ms ile 150ms aras\u0131<\/td>\nEmir ger\u00e7ekle\u015fti\u011finde Broker B zaten g\u00fcncellenmi\u015ftir \u2014 zarar veya requote<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Kirlenme oran\u0131 sorunu<\/h3>\n

Filtreleme olmadan, tipik bir iki brokerl\u0131 kurulumda gecikme kaynakl\u0131 sinyallerin ger\u00e7ek LP fark\u0131 sinyallerine oran\u0131 %60\u201380 gecikme kaynakl\u0131 yapay sinyal \u015feklindedir. Piyasa olaylar\u0131 \u2014 fiyat tikleri, ekonomik veri a\u00e7\u0131klamalar\u0131, likidite kaymalar\u0131 \u2014 s\u00fcrekli ger\u00e7ekle\u015fir ve her biri, daha h\u0131zl\u0131 broker\u0131n fiyat ak\u0131\u015f\u0131 \u00f6nce g\u00fcncellendi\u011fi i\u00e7in ge\u00e7ici, gecikme kaynakl\u0131 bir fark yarat\u0131r. Bu oran b\u00fcy\u00fck \u00f6l\u00e7\u00fcde iki broker aras\u0131ndaki fiyat ak\u0131\u015f\u0131 h\u0131z fark\u0131na ba\u011fl\u0131d\u0131r: fiyat ak\u0131\u015flar\u0131ndan biri di\u011ferinden s\u00fcrekli 100ms daha h\u0131zl\u0131 olan bir broker \u00e7ifti, birbirine 10ms yak\u0131nl\u0131ktaki bir \u00e7ifte g\u00f6re \u00e7ok daha fazla yapay sinyal \u00fcretir.<\/p>\n

Gecikme kaynakl\u0131 sinyaller neden toksik emir ak\u0131\u015f\u0131 yarat\u0131r<\/h2>\n

Gecikme kaynakl\u0131 yapay sinyallerle i\u015flem yapmak, tekil i\u015flem d\u00fczeyinde k\u00e2rs\u0131zd\u0131r. Ancak sorun bununla s\u0131n\u0131rl\u0131 de\u011fildir \u2014 brokerlar\u0131n teknik a\u00e7\u0131dan toksik olarak s\u0131n\u0131fland\u0131rd\u0131\u011f\u0131 belirli bir emir ak\u0131\u015f\u0131 \u00f6r\u00fcnt\u00fcs\u00fc olu\u015fturur.<\/p>\n

\n
\n
\u00d6R\u00dcNT\u00dc 1<\/div>\n

Fiyat hareketi anlar\u0131nda verilen emirler<\/h3>\n

Gecikme kaynakl\u0131 sinyaller tam olarak bir fiyat olay\u0131 ger\u00e7ekle\u015fti\u011finde tetiklenir \u2014 \u00e7\u00fcnk\u00fc h\u0131zl\u0131 broker\u0131n fiyat ak\u0131\u015f\u0131 tam o anda g\u00fcncellenirken yava\u015f broker\u0131nki hen\u00fcz g\u00fcncellenmemi\u015ftir. Filtrelenmemi\u015f hedge arbitraj\u0131, emirlerinin orant\u0131s\u0131z bir b\u00f6l\u00fcm\u00fcn\u00fc tam da fiyat hareketi olaylar\u0131n\u0131n ya\u015fand\u0131\u011f\u0131 anlarda verir. Emir zaman damgalar\u0131n\u0131 fiyat olay\u0131 zaman damgalar\u0131yla ili\u015fkilendiren broker risk sistemleri, bu yo\u011funla\u015fmay\u0131 teknik ak\u0131\u015f olarak tan\u0131mlar.<\/p>\n<\/div>\n

\n
\u00d6R\u00dcNT\u00dc 2<\/div>\n

Y\u00fcksek ters se\u00e7im (adverse selection) oran\u0131<\/h3>\n

Gecikme kaynakl\u0131 i\u015flemler, birka\u00e7 y\u00fcz milisaniye sonraki piyasaya g\u00f6re s\u00fcrekli olarak elveri\u015fsiz fiyatlardan ger\u00e7ekle\u015fir. Broker risk sistemleri her emir ger\u00e7ekle\u015fmesinden sonraki k\u0131sa vadeli fiyat seyrini takip eder. Ger\u00e7ekle\u015fmelerin hemen ard\u0131ndan ters y\u00f6nde fiyat hareketinin geldi\u011fi y\u00fcksek bir oran \u2014 piyasan\u0131n tam emir y\u00f6n\u00fcn\u00fcn aksine hareket etti\u011fi anda ula\u015fan emirlerin imzas\u0131 \u2014 hesab\u0131n ters se\u00e7im incelemesine al\u0131nmas\u0131na yol a\u00e7ar.<\/p>\n<\/div>\n

\n
\u00d6R\u00dcNT\u00dc 3<\/div>\n

Volatiliteyle ters orant\u0131l\u0131 K&Z<\/h3>\n

Gecikme kaynakl\u0131 yapay sinyallerle i\u015flem yapan bir sistem, d\u00fc\u015f\u00fck volatilite d\u00f6nemlerinde k\u00e2rl\u0131 (ger\u00e7ek LP farklar\u0131 bask\u0131nd\u0131r), y\u00fcksek volatilite d\u00f6nemlerinde ise k\u00e2rs\u0131zd\u0131r (yapay sinyaller \u00e7o\u011fal\u0131r). K&Z ile volatilite aras\u0131ndaki bu ters korelasyon \u2014 me\u015fru teknik i\u015flemlerle ba\u011fda\u015fmaz \u2014 tan\u0131nabilir bir toksik ak\u0131\u015f imzas\u0131d\u0131r.<\/p>\n<\/div>\n

\n
\u00d6R\u00dcNT\u00dc 4<\/div>\n

Asimetrik pozisyon tutma s\u00fcresi da\u011f\u0131l\u0131m\u0131<\/h3>\n

Gecikme kaynakl\u0131 bir i\u015flem yanl\u0131\u015f fiyattan ger\u00e7ekle\u015fti\u011finde sistem birka\u00e7 saniye i\u00e7inde pozisyondan \u00e7\u0131kar. K\u00e2rl\u0131 i\u015flemlerin daha uzun tutuldu\u011fu, zararl\u0131 i\u015flemlerin ise h\u0131zla kapat\u0131ld\u0131\u011f\u0131 bu \u00f6r\u00fcnt\u00fc, risk sistemlerinin otomatik teknik istismar olarak i\u015faretledi\u011fi asimetrik bir tutma s\u00fcresi da\u011f\u0131l\u0131m\u0131 \u00fcretir.<\/p>\n<\/div>\n<\/div>\n

\n

S\u0131n\u0131fland\u0131rman\u0131n sonucu<\/h3>\n

Toksik ak\u0131\u015f olarak s\u0131n\u0131fland\u0131r\u0131lan hesap k\u00f6t\u00fcle\u015ftirilmi\u015f emir ger\u00e7ekle\u015ftirme ile kar\u015f\u0131la\u015f\u0131r \u2014 requote’lar, spread geni\u015flemesi veya broker\u0131n kendi defterine dahili y\u00f6nlendirme. Bu durum k\u00e2rl\u0131 ger\u00e7ek LP fark\u0131 i\u015flemlerini de olumsuz etkiler.<\/strong> Filtre sorunu yaln\u0131zca tekil i\u015flemlerin k\u00e2rl\u0131l\u0131\u011f\u0131yla ilgili de\u011fildir \u2014 hesab\u0131n arbitraj i\u00e7in kullan\u0131labilir kal\u0131p kalmayaca\u011f\u0131n\u0131 belirler.<\/p>\n<\/div>\n

Basit filtre: difference-to-open e\u015fi\u011fi<\/h2>\n
\n

En basit filtre, alt\u0131nda hi\u00e7bir emir verilmeyen asgari bir fiyat fark\u0131 e\u015fi\u011fidir \u2014 difference-to-open.<\/strong> Gecikme kaynakl\u0131 farklar ortalamada ger\u00e7ek LP farklar\u0131ndan daha k\u00fc\u00e7\u00fck olma e\u011filimindedir: fiyat ak\u0131\u015f\u0131 gecikmesi kaynakl\u0131 yapay farklar yaln\u0131zca ayn\u0131 olay\u0131 alan iki ak\u0131\u015f aras\u0131ndaki h\u0131z fark\u0131n\u0131 yans\u0131t\u0131rken, ger\u00e7ek LP farklar\u0131 daha b\u00fcy\u00fck ve daha kal\u0131c\u0131 olabilen fiili LP d\u00fczeyindeki fiyatlama farklar\u0131n\u0131 yans\u0131t\u0131r. Asgari bir e\u015fik belirlemek, daha b\u00fcy\u00fck ger\u00e7ek sinyalleri korurken daha k\u00fc\u00e7\u00fck gecikme kaynakl\u0131 sinyalleri eler.<\/p>\n<\/div>\n

E\u015fi\u011fin kalibrasyonu<\/h3>\n\n\n\n\n\n\n\n\n
Ad\u0131m<\/th>\n\u0130\u015flem<\/th>\n\u00c7\u0131kt\u0131<\/th>\n<\/tr>\n<\/thead>\n
Ad\u0131m 1<\/td>\nSistemi 5\u201310 i\u015flem g\u00fcn\u00fc boyunca yaln\u0131zca izleme modunda \u00e7al\u0131\u015ft\u0131r\u0131n. Tespit edilen her fark\u0131 \u015fu bilgilerle kaydedin: zaman damgas\u0131, b\u00fcy\u00fckl\u00fck (pip), kapanmadan \u00f6nceki s\u00fcre ve tespitten sonraki 200ms i\u00e7indeki fiyat hareketi.<\/td>\nHam sinyal veri seti<\/td>\n<\/tr>\n
Ad\u0131m 2<\/td>\nHer sinyali sonradan s\u0131n\u0131fland\u0131r\u0131n: ger\u00e7ek (150ms’den uzun s\u00fcrer, referans fiyat ak\u0131\u015f\u0131 durgun) veya gecikme kaynakl\u0131 (100ms’den k\u0131sa s\u00fcrede kapan\u0131r, \u00f6ncesinde y\u00f6nl\u00fc bir fiyat olay\u0131 vard\u0131r).<\/td>\nS\u0131n\u0131fland\u0131r\u0131lm\u0131\u015f veri seti<\/td>\n<\/tr>\n
Ad\u0131m 3<\/td>\nGer\u00e7ek ve gecikme kaynakl\u0131 sinyallerin b\u00fcy\u00fckl\u00fck da\u011f\u0131l\u0131m\u0131n\u0131 grafi\u011fe d\u00f6k\u00fcn. \u00dczerindeki sinyallerin %80+’s\u0131n\u0131n ger\u00e7ek LP fark\u0131 oldu\u011fu e\u015fi\u011fi bulun.<\/td>\nOptimum e\u015fik de\u011feri<\/td>\n<\/tr>\n
Ad\u0131m 4<\/td>\nDifference-to-open de\u011ferini belirlenen e\u015fi\u011fe ayarlay\u0131n. \u00dc\u00e7 ayda bir yeniden kalibre edin \u2014 optimum e\u015fik, brokerlar\u0131n LP ili\u015fkileri ve piyasa volatilite rejimleri de\u011fi\u015ftik\u00e7e kayar.<\/td>\nDevreye al\u0131nan filtre parametresi<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

\u00dc\u00e7 yap\u0131sal s\u0131n\u0131rlama<\/h3>\n
\n
\n
SINIRLAMA 1<\/div>\n

B\u00fcy\u00fck gecikme olaylar\u0131 filtreden ge\u00e7er<\/h3>\n

Y\u00fcksek volatilite d\u00f6nemleri b\u00fcy\u00fck fiyat hareketleri yarat\u0131r. 10 pip’lik bir hareket s\u0131ras\u0131nda olu\u015fan fiyat ak\u0131\u015f\u0131 gecikmesi kaynakl\u0131 yapay fark, brokerlar aras\u0131nda 4\u20135 pip’lik bir fark olarak g\u00f6r\u00fcnebilir. 3 pip’lik bir e\u015fik bunu yakalayamaz. B\u00fcy\u00fck gecikme olaylar\u0131 en zararl\u0131 olanlard\u0131r \u2014 en b\u00fcy\u00fck kay\u0131plar ve en g\u00f6r\u00fcn\u00fcr toksik ak\u0131\u015f \u00f6r\u00fcnt\u00fcs\u00fc.<\/p>\n<\/div>\n

\n
SINIRLAMA 2<\/div>\n

Statik e\u015fik, dinamik piyasa<\/h3>\n

Optimum e\u015fik piyasa ko\u015fullar\u0131na g\u00f6re de\u011fi\u015fir. Asya seans\u0131n\u0131n sakin d\u00f6nemlerinde 1,5 pip’lik bir e\u015fik yeterli filtreleme sa\u011flayabilir. Londra a\u00e7\u0131l\u0131\u015f\u0131nda ise ayn\u0131 e\u015fik, gecikme kaynakl\u0131 yapay sinyallerin y\u00fcksek bir k\u0131sm\u0131n\u0131 ge\u00e7irir. Sabit bir e\u015fik her zaman bir uzla\u015fmad\u0131r.<\/p>\n<\/div>\n

\n
SINIRLAMA 3<\/div>\n

Y\u00f6n fark\u0131ndal\u0131\u011f\u0131 yok<\/h3>\n

Basit filtre, sinyalin tespit edildi\u011fi andaki piyasa ba\u011flam\u0131ndan ba\u011f\u0131ms\u0131z olarak e\u015fi\u011fin \u00fczerindeki t\u00fcm farklar\u0131 ayn\u0131 \u015fekilde de\u011ferlendirir. 3 pip’lik ger\u00e7ek bir LP fark\u0131n\u0131, h\u0131zl\u0131 bir fiyat olay\u0131n\u0131n \u00fcretti\u011fi 3 pip’lik gecikme kaynakl\u0131 yapay farktan ay\u0131rt edemez.<\/p>\n<\/div>\n<\/div>\n

Geli\u015fmi\u015f filtre: h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 referans mimarisi<\/h2>\n
\n

H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi, tespit edilen her fark\u0131 s\u0131n\u0131fland\u0131rma mekanizmas\u0131 olarak her iki emir ger\u00e7ekle\u015ftirme broker\u0131ndan da daha h\u0131zl\u0131 olan \u00fc\u00e7\u00fcnc\u00fc, ba\u011f\u0131ms\u0131z bir veri kayna\u011f\u0131 ekler.<\/strong> Sistem yaln\u0131zca b\u00fcy\u00fckl\u00fc\u011fe g\u00f6re filtrelemek yerine, tespit edilen fark\u0131n referans ak\u0131\u015ftaki bir fiyat hareketi olay\u0131yla \u00e7ak\u0131\u015f\u0131p \u00e7ak\u0131\u015fmad\u0131\u011f\u0131n\u0131 belirlemek i\u00e7in h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131n\u0131 kullan\u0131r. \u00c7ak\u0131\u015f\u0131yorsa sinyal gecikme kaynakl\u0131d\u0131r ve at\u0131l\u0131r. Referans ak\u0131\u015f durgunsa fark ger\u00e7ek bir LP fark\u0131 olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r ve emir ger\u00e7ekle\u015ftirme devam eder.<\/p>\n<\/div>\n

Her sinyal tespiti olay\u0131ndaki s\u0131n\u0131fland\u0131rma mant\u0131\u011f\u0131:<\/p>\n

\n

Sinyal, T\u00dcM ko\u015fullar do\u011fruysa GER\u00c7EK LP FARKI olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r:<\/p>\n

(1) FastFeed.RateOfChange(last 100ms) < MovementThreshold
\n\u2014 Referans fiyat ak\u0131\u015f\u0131 \u00f6nceki pencerede kayda de\u011fer bi\u00e7imde hareket etmedi
\n\u2014 \u015eu anda piyasada yay\u0131lan y\u00f6nl\u00fc bir fiyat olay\u0131 yok<\/p>\n

(2) FastFeed.LastTickAge < StalenessThreshold
\n\u2014 Referans fiyat ak\u0131\u015f\u0131 aktif olarak g\u00fcncelleniyor (bayat veya ba\u011flant\u0131s\u0131 kopmu\u015f de\u011fil)
\n\u2014 S\u0131n\u0131fland\u0131rma kontrol\u00fcn\u00fcn canl\u0131 veriye dayanmas\u0131n\u0131 sa\u011flar<\/p>\n

(3) Discrepancy.Persistence > MinPersistenceWindow
\n\u2014 Fark en az N milisaniyedir mevcut
\n\u2014 Ger\u00e7ek LP farklar\u0131 kal\u0131c\u0131d\u0131r; yapay gecikme farklar\u0131 30\u2013100ms i\u00e7inde kapan\u0131r<\/p>\n

HERHANG\u0130 B\u0130R ko\u015ful sa\u011flanmazsa sinyal GEC\u0130KME KAYNAKLI olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r \u2192 emir iptal edilir.<\/p>\n<\/div>\n

Ko\u015ful 1: H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131ndaki de\u011fi\u015fim oran\u0131<\/h3>\n

Bu, birincil s\u0131n\u0131fland\u0131rma ko\u015fuludur. H\u0131zl\u0131 referans fiyat ak\u0131\u015f\u0131 her iki emir ger\u00e7ekle\u015ftirme broker\u0131ndan da daha h\u0131zl\u0131d\u0131r \u2014 k\u00fcresel piyasa olaylar\u0131n\u0131, iki broker\u0131n kotasyonlar\u0131 g\u00fcncellenmeden \u00f6nce al\u0131r. Referans ak\u0131\u015f, sinyal tespitinden \u00f6nceki pencerede kayda de\u011fer bir fiyat hareketi g\u00f6steriyorsa, bu hareket g\u00f6r\u00fcnen fark\u0131n olas\u0131 nedenidir: emir ger\u00e7ekle\u015ftirme brokerlar\u0131ndan birinin fiyat ak\u0131\u015f\u0131 olaydan g\u00fcncellenmi\u015f, di\u011feri ise hen\u00fcz g\u00fcncellenmemi\u015ftir.<\/p>\n\n\n\n\n\n\n\n
Parametre<\/th>\nTipik de\u011fer<\/th>\nNot<\/th>\n<\/tr>\n<\/thead>\n
MovementThreshold<\/td>\nMaj\u00f6r pariteler i\u00e7in 100ms ba\u015f\u0131na 0,5\u20131,5 pip; volatil pariteler i\u00e7in 1,5\u20133,0 pip<\/td>\n\u0130lgili enstr\u00fcman i\u00e7in normal ve olay ko\u015fullar\u0131nda 100ms ba\u015f\u0131na tipik pip hareketinin \u00f6l\u00e7\u00fclmesiyle belirlenir<\/td>\n<\/tr>\n
Geriye bak\u0131\u015f penceresi<\/td>\n50\u2013200ms<\/td>\nHareket olay\u0131n\u0131 yakalamal\u0131, ancak bir hareket olay\u0131n\u0131 gecikmeyle izleyen ger\u00e7ek farklar\u0131 maskelememelidir. \u00d6l\u00e7\u00fclen fiyat ak\u0131\u015f\u0131 gecikme fark\u0131ndan kalibre edilir.<\/td>\n<\/tr>\n
Bayatl\u0131k e\u015fi\u011fi<\/td>\n500ms\u20132.000ms<\/td>\nReferans fiyat ak\u0131\u015f\u0131 bu pencere i\u00e7inde g\u00fcncellenmediyse tak\u0131lm\u0131\u015f olabilir. Bayatl\u0131k olaylar\u0131 s\u0131ras\u0131ndaki sinyallere ihtiyatl\u0131 yakla\u015f\u0131n \u2014 ak\u0131\u015f yeniden ba\u015flayana kadar at\u0131n veya bekletin.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi neden niteliksel olarak \u00fcst\u00fcnd\u00fcr<\/h3>\n
\n
\n
AVANTAJ 1<\/div>\n

B\u00fcy\u00fck gecikme olaylar\u0131n\u0131 yakalar<\/h3>\n

B\u00fcy\u00fck bir fiyat olay\u0131 s\u0131ras\u0131ndaki 5 pip’lik yapay gecikme fark\u0131, b\u00fcy\u00fckl\u00fc\u011f\u00fcnden ba\u011f\u0131ms\u0131z olarak h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi taraf\u0131ndan gecikme kaynakl\u0131 olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r. Basit e\u015fik 5 pip’in alt\u0131na ayarlanm\u0131\u015fsa bu fark\u0131 ge\u00e7irir. H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi, fark b\u00fcy\u00fckl\u00fc\u011f\u00fcne de\u011fil piyasa durumuna g\u00f6re s\u0131n\u0131fland\u0131rma yapar.<\/p>\n<\/div>\n

\n
AVANTAJ 2<\/div>\n

Volatiliteye dinamik olarak uyum sa\u011flar<\/h3>\n

Sakin d\u00f6nemlerde referans ak\u0131\u015f d\u00fc\u015f\u00fck de\u011fi\u015fim oran\u0131 g\u00f6sterir ve s\u0131n\u0131fland\u0131rma nadiren tetiklenir. Volatil d\u00f6nemlerde referans ak\u0131\u015f aktiftir ve gecikme kaynakl\u0131 yapay sinyalleri s\u00fcrekli olarak s\u0131n\u0131fland\u0131r\u0131r. Filtre, manuel ayar gerektirmeden y\u00fcksek riskli d\u00f6nemlerde kendili\u011finden s\u0131k\u0131la\u015f\u0131r.<\/p>\n<\/div>\n

\n
AVANTAJ 3<\/div>\n

K\u00fc\u00e7\u00fck ger\u00e7ek sinyalleri korur<\/h3>\n

Basit e\u015fi\u011fin engelleyece\u011fi 0,8 pip’lik ger\u00e7ek bir LP fark\u0131, referans ak\u0131\u015f durgunsa h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresinden ge\u00e7er. Filtre, k\u00fc\u00e7\u00fck gecikme sinyallerinden ka\u00e7\u0131nmak i\u00e7in k\u00fc\u00e7\u00fck ger\u00e7ek sinyalleri feda etmez \u2014 her sinyal piyasa ba\u011flam\u0131na g\u00f6re ba\u011f\u0131ms\u0131z olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r.<\/p>\n<\/div>\n

\n
AVANTAJ 4<\/div>\n

Gecikmenin neredeyse tamamen ortadan kald\u0131r\u0131lmas\u0131<\/h3>\n

\u0130yi kalibre edilmi\u015f kurulumlarda h\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi, gecikme kaynakl\u0131 kirlenmeyi filtrelenmemi\u015f sinyallerin %60\u201380’inden %5’in alt\u0131na indirir. Kalan kirlenme, bir fiyat olay\u0131n\u0131n de\u011fi\u015fim oran\u0131 e\u015fi\u011fini tetiklemeyecek kadar yava\u015f, ancak fiyat ak\u0131\u015f\u0131 fark\u0131 yaratacak kadar h\u0131zl\u0131 oldu\u011fu u\u00e7 durumlardan olu\u015fur.<\/p>\n<\/div>\n<\/div>\n

Filtre kar\u015f\u0131la\u015ft\u0131rmas\u0131<\/h2>\n\n\n\n\n\n\n\n\n
Filtre<\/th>\nGecikme eleme<\/th>\nGer\u00e7ek sinyal koruma<\/th>\nKarma\u015f\u0131kl\u0131k<\/th>\nEn uygun kullan\u0131m<\/th>\n<\/tr>\n<\/thead>\n
Filtre yok<\/td>\n%0 \u2014 hepsi ge\u00e7er<\/td>\n100%<\/td>\nYok<\/td>\nUygulanabilir de\u011fil \u2014 toksik ak\u0131\u015f ka\u00e7\u0131n\u0131lmaz<\/td>\n<\/tr>\n
Yaln\u0131zca difference-to-open<\/td>\n40\u201360%<\/td>\n80\u201390%<\/td>\nD\u00fc\u015f\u00fck<\/td>\n\u0130lk kurulumlar, fiyat ak\u0131\u015f\u0131 h\u0131zlar\u0131 birbirine yak\u0131n brokerlar<\/td>\n<\/tr>\n
Difference-to-open + kal\u0131c\u0131l\u0131k<\/td>\n60\u201375%<\/td>\n75\u201385%<\/td>\nD\u00fc\u015f\u00fck<\/td>\n\u0130yile\u015ftirilmi\u015f temel seviye, b\u00fcy\u00fck gecikme olaylar\u0131n\u0131 yine ka\u00e7\u0131r\u0131r<\/td>\n<\/tr>\n
H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi (tam)<\/td>\n90\u201397%<\/td>\n85\u201395%<\/td>\nY\u00fcksek<\/td>\nCanl\u0131 arbitraj, brokerlar\u0131n d\u00fc\u015fmanca yakla\u015ft\u0131\u011f\u0131 ortamlar<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\n

H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi i\u00e7in kalibrasyon s\u0131ras\u0131<\/h3>\n
    \n
  • A\u015fama 1:<\/strong> Sistemi 5\u201310 g\u00fcn boyunca kay\u0131t modunda \u00e7al\u0131\u015ft\u0131r\u0131n. Tespit edilen her fark\u0131 t\u00fcm ba\u011flam\u0131yla kaydedin: t\u00fcm fiyat ak\u0131\u015flar\u0131n\u0131n durumu, fark b\u00fcy\u00fckl\u00fc\u011f\u00fc, kal\u0131c\u0131l\u0131k s\u00fcresi ve \u00f6nceki 100ms ve 200ms pencerelerindeki referans ak\u0131\u015f de\u011fi\u015fim oran\u0131.<\/li>\n
  • A\u015fama 2:<\/strong> Kaydedilen her sinyali sonradan ger\u00e7ek (150ms’den uzun s\u00fcrer, referans ak\u0131\u015f durgun) veya gecikme kaynakl\u0131 (100ms’den k\u0131sa s\u00fcrede kapan\u0131r, tespit an\u0131nda referans ak\u0131\u015f aktif) olarak s\u0131n\u0131fland\u0131r\u0131n.<\/li>\n
  • A\u015fama 3:<\/strong> MovementThreshold, geriye bak\u0131\u015f penceresi ve MinPersistenceWindow parametrelerinin farkl\u0131 kombinasyonlar\u0131n\u0131 s\u0131n\u0131fland\u0131r\u0131lm\u0131\u015f veri setine kar\u015f\u0131 test edin. Hedef: ger\u00e7ek sinyal koruma %85’in \u00fczerinde, gecikme ge\u00e7i\u015fi %8’in alt\u0131nda.<\/li>\n
  • A\u015fama 4:<\/strong> Optimize edilmi\u015f parametrelerle minimum lot b\u00fcy\u00fckl\u00fc\u011f\u00fcnde devreye al\u0131n. \u0130lk 10 i\u015flem g\u00fcn\u00fc boyunca sinyal t\u00fcr\u00fcne g\u00f6re K&Z da\u011f\u0131l\u0131m\u0131n\u0131 izleyin. \u00dc\u00e7 ayda bir veya \u00f6nemli piyasa rejimi de\u011fi\u015fikliklerinden sonra yeniden kalibre edin.<\/li>\n<\/ul>\n<\/div>\n

    S\u0131k\u00e7a Sorulan Sorular<\/h2>\n
    \n
    Ger\u00e7ek bir LP fark\u0131 ile gecikme kaynakl\u0131 bir fark aras\u0131ndaki fark nedir?<\/div>\n
    \n

    Ger\u00e7ek bir LP fark\u0131, iki broker\u0131n likidite sa\u011flay\u0131c\u0131lar\u0131n\u0131n ayn\u0131 enstr\u00fcman\u0131 farkl\u0131 fiyatlamas\u0131ndan kaynaklan\u0131r \u2014 her LP’nin piyasay\u0131 modelleme veya envanter y\u00f6netme bi\u00e7imindeki yap\u0131sal bir farkl\u0131l\u0131k. Gecikme kaynakl\u0131 bir fark ise brokerlardan birinin fiyat ak\u0131\u015f\u0131n\u0131n bir piyasa olay\u0131n\u0131 di\u011ferinden daha h\u0131zl\u0131 almas\u0131ndan kaynaklan\u0131r. Ger\u00e7ek farklar i\u015flem yap\u0131labilirdir ve 200ms ile 2.000ms aras\u0131nda s\u00fcrer. Yapay gecikme farklar\u0131, daha yava\u015f ak\u0131\u015f yeti\u015ftik\u00e7e 30\u2013150ms i\u00e7inde kapan\u0131r \u2014 genellikle emirler daha yava\u015f brokera ula\u015fmadan \u00f6nce.<\/p>\n<\/div>\n

    Difference-to-open nedir ve nas\u0131l yard\u0131mc\u0131 olur?<\/div>\n
    \n

    Difference-to-open, herhangi bir emir verilmeden \u00f6nce iki broker kotasyonu aras\u0131nda bulunmas\u0131 gereken, pip cinsinden ayarlanabilir asgari fiyat fark\u0131d\u0131r. Fiyat ak\u0131\u015f\u0131 gecikmesi kaynakl\u0131 yapay farklar\u0131n ortalamada ger\u00e7ek LP farklar\u0131ndan daha k\u00fc\u00e7\u00fck olma e\u011filiminden yararlanarak gecikme kaynakl\u0131 sinyalleri filtreler. K\u00fc\u00e7\u00fck gecikme sinyallerini elemede etkilidir, ancak h\u0131zl\u0131 bir fiyat hareketinin farkl\u0131 g\u00fcncelleme h\u0131zlar\u0131na sahip ak\u0131\u015flar aras\u0131nda b\u00fcy\u00fck bir g\u00f6r\u00fcn\u00fcr fark yaratt\u0131\u011f\u0131 y\u00fcksek volatilite d\u00f6nemlerindeki b\u00fcy\u00fck gecikme olaylar\u0131n\u0131 yakalayamaz.<\/p>\n<\/div>\n

    H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi nedir ve nas\u0131l \u00e7al\u0131\u015f\u0131r?<\/div>\n
    \n

    H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi, emir vermeden \u00f6nce tespit edilen her fark\u0131 s\u0131n\u0131fland\u0131rmak i\u00e7in her iki emir ger\u00e7ekle\u015ftirme broker\u0131ndan da daha h\u0131zl\u0131 olan \u00fc\u00e7\u00fcnc\u00fc bir referans fiyat ak\u0131\u015f\u0131 kullan\u0131r. Referans ak\u0131\u015f, sinyal tespitinden \u00f6nceki pencerede kayda de\u011fer bir fiyat hareketi g\u00f6steriyorsa fark gecikme kaynakl\u0131 olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r ve at\u0131l\u0131r. Referans ak\u0131\u015f durgunsa fark potansiyel bir ger\u00e7ek LP fark\u0131 olarak s\u0131n\u0131fland\u0131r\u0131l\u0131r ve emir ger\u00e7ekle\u015ftirme devam eder. Bu y\u00f6ntem sinyalleri yaln\u0131zca b\u00fcy\u00fckl\u00fc\u011fe g\u00f6re de\u011fil piyasa durumuna g\u00f6re s\u0131n\u0131fland\u0131r\u0131r ve basit e\u015fi\u011fin ka\u00e7\u0131rd\u0131\u011f\u0131 b\u00fcy\u00fck gecikme olaylar\u0131n\u0131 yakalar.<\/p>\n<\/div>\n

    Filtrelenmemi\u015f hedge arbitraj\u0131 neden toksik emir ak\u0131\u015f\u0131 yarat\u0131r?<\/div>\n
    \n

    Gecikme kaynakl\u0131 sinyaller, tam olarak fiyat hareketi olaylar\u0131n\u0131n ya\u015fand\u0131\u011f\u0131 anda emir \u00fcretir. Broker risk sistemleri bu yo\u011funla\u015fmay\u0131 teknik bir ak\u0131\u015f \u00f6r\u00fcnt\u00fcs\u00fc olarak tespit eder. Ayr\u0131ca gecikme kaynakl\u0131 i\u015flemler s\u00fcrekli olarak elveri\u015fsiz ger\u00e7ekle\u015fmeler ve k\u00fc\u00e7\u00fck zararlarla h\u0131zl\u0131 \u00e7\u0131k\u0131\u015flar \u00fcretir; bu da hesab\u0131 teknik istismar olarak daha da belirginle\u015ftiren asimetrik bir tutma s\u00fcresi ve K&Z da\u011f\u0131l\u0131m\u0131 olu\u015fturur. S\u0131n\u0131fland\u0131r\u0131ld\u0131ktan sonra hesap, k\u00e2rl\u0131 ger\u00e7ek LP fark\u0131 i\u015flemlerini de etkileyen k\u00f6t\u00fcle\u015ftirilmi\u015f emir ger\u00e7ekle\u015ftirme ile kar\u015f\u0131la\u015f\u0131r.<\/p>\n<\/div>\n

    H\u0131zl\u0131 fiyat ak\u0131\u015f\u0131 filtresi i\u00e7in ne gereklidir?<\/div>\n
    \n

    K\u00fcresel piyasa olaylar\u0131n\u0131 almada her iki emir ger\u00e7ekle\u015ftirme broker\u0131ndan da kan\u0131tlanabilir \u015fekilde daha h\u0131zl\u0131 olan \u00fc\u00e7\u00fcnc\u00fc bir fiyat veri kayna\u011f\u0131. Bu, \u00f6zel bir piyasa verisi sa\u011flay\u0131c\u0131s\u0131, co-located bir fiyat ak\u0131\u015f\u0131 hizmeti veya yaln\u0131zca referans ak\u0131\u015f olarak kullan\u0131lan (\u00fczerinden emir verilmeyen) \u00fc\u00e7\u00fcnc\u00fc bir broker hesab\u0131 olabilir. Referans ak\u0131\u015f, soket al\u0131m\u0131nda donan\u0131m d\u00fczeyinde zaman damgalar\u0131yla ayr\u0131 bir izleme i\u015f par\u00e7ac\u0131\u011f\u0131nda \u00e7al\u0131\u015f\u0131r. Referans ak\u0131\u015f \u00fczerinden asla emir verilmez \u2014 yaln\u0131zca bir s\u0131n\u0131fland\u0131rma girdisidir.<\/p>\n<\/div>\n

    Filtre parametrelerinin ne s\u0131kl\u0131kla yeniden kalibre edilmesi gerekir?<\/div>\n
    \n

    En az \u00fc\u00e7 ayda bir. Optimum parametreler, brokerlar\u0131n LP ili\u015fkileri geli\u015ftik\u00e7e, fiyat ak\u0131\u015f\u0131 altyap\u0131s\u0131 de\u011fi\u015ftik\u00e7e ve piyasa volatilite rejimleri kayd\u0131k\u00e7a de\u011fi\u015fir. \u015eu durumlarda derhal yeniden kalibrasyon gerekir: bir broker LP’sini veya fiyatlama motorunu de\u011fi\u015ftirdi\u011finde; VPS sa\u011flay\u0131c\u0131s\u0131 veya a\u011f y\u00f6nlendirmesindeki de\u011fi\u015fiklikler fiyat ak\u0131\u015f\u0131 gecikmelerini etkiledi\u011finde; ya da enstr\u00fcman volatilitesinde kal\u0131c\u0131 de\u011fi\u015fiklikler oldu\u011funda. SharpTrader’\u0131n sinyal baz\u0131nda at\u0131f kayd\u0131, ayr\u0131 bir \u00e7evrimd\u0131\u015f\u0131 moda gerek kalmadan s\u00fcrekli kalibrasyon i\u00e7in gereken veriyi sa\u011flar.<\/p>\n<\/div>\n<\/div>\n

    \n

    SharpTrader Pro \u2014 H\u0131zl\u0131 Fiyat Ak\u0131\u015f\u0131 Filtrelemeli Hedge Arbitraj\u0131<\/h2>\n

    Difference-to-open \u00b7 H\u0131zl\u0131 referans fiyat ak\u0131\u015f\u0131 filtresi \u00b7 Sinyal baz\u0131nda at\u0131f kayd\u0131 \u00b7 FIX API uyumlu \u00b7 25 y\u0131ll\u0131k arbitraj geli\u015ftirme deneyimi<\/p>\n

    SharpTrader Pro’yu inceleyin \u2192<\/a>
    \n    Gecikme Arbitraj\u0131 \u2192<\/a><\/p>\n<\/div>\n<\/div>\n