%% Initialize
clear all;
startDate = datenum('2005-08-25');
endDate = datenum('2010-10-20');
dateRange = busdays(startDate, endDate);

%% Load SPY data
csvFile = fopen('data/SPY.csv', 'r');
csvFields = textscan(csvFile, '%s %f %f %f %f %f %f',-1,'delimiter',',','HeaderLines',1);
fclose(csvFile);
ftsSPY = fints(datenum(csvFields{1},'yyyy-mm-dd'), [csvFields{2} csvFields{3}, csvFields{4} csvFields{5} csvFields{6} csvFields{7}], {'Open','High','Low','Close','Volume','Adj'});
ftsSPY = ftsSPY(ismember(ftsSPY.dates, dateRange));

%% Load POMO data
csvFile = fopen('data/pomo_20050825_20101025.csv', 'r');
pomoData = textscan(csvFile, '%s %s %s %s %s %s %s %f %f %f', -1, 'delimiter', ',', 'HeaderLines', 1);
fclose(csvFile);
directionVec = strcmpi(pomoData{2}, 'P');
directionVec(directionVec==0) = -1;
ftsPOMO = fints(datenum(pomoData{1}), [pomoData{9}, pomoData{10}, directionVec, pomoData{8}], {'Accepted','Submitted','Direction','NumIssues'});
ftsPOMO = ftsPOMO(ismember(ftsPOMO.dates, dateRange));
ftsPOMO = todaily(ftsPOMO);

%% Double-check dates
if length(ftsSPY) ~= length(ftsPOMO)
    error('Date vectors do not match in length between SPY and POMO');
else
    if sum(ftsSPY.dates ~= ftsPOMO.dates) > 0
        error('Date vectors do not match in values between SPY and POMO');
    end
end

%% Run the regression on submitted/accepted ratio
% Close - open intraday return
spyIntraDayReturn = log(fts2mat(ftsSPY.Close)) - log(fts2mat(ftsSPY.Open));

% Log total dollars traded, as approximated by open-close midpoint
spyDolVol = log(0.5 * fts2mat(ftsSPY.Volume) .* (fts2mat(ftsSPY.Open) + fts2mat(ftsSPY.Close)));

% Submit/accept ratio
pomoSubmitAcceptRatio = fts2mat(ftsPOMO.Direction) .* fts2mat(ftsPOMO.Accepted) ./ fts2mat(ftsPOMO.Submitted);

pomoTopThird = find(pomoSubmitAcceptRatio > prctile(pomoSubmitAcceptRatio, 66));
pomoBottomThird = find(pomoSubmitAcceptRatio < prctile(pomoSubmitAcceptRatio, 33));
pomoMiddleThird = find((pomoSubmitAcceptRatio >= prctile(pomoSubmitAcceptRatio, 33)) .* (pomoSubmitAcceptRatio <= prctile(pomoSubmitAcceptRatio, 66)));
noPomo = find(isnan(pomoSubmitAcceptRatio));

figure;
hold all;
disp('# Submit/Accept');
scatter(pomoSubmitAcceptRatio(pomoBottomThird), spyIntraDayReturn(pomoBottomThird), 'or');
fprintf('Bottom: %f\n', sum(spyIntraDayReturn(pomoBottomThird)));
scatter(pomoSubmitAcceptRatio(pomoMiddleThird), spyIntraDayReturn(pomoMiddleThird), 'xb');
fprintf('Middle: %f\n', sum(spyIntraDayReturn(pomoMiddleThird)));
scatter(pomoSubmitAcceptRatio(pomoTopThird), spyIntraDayReturn(pomoTopThird), '.g');
fprintf('Top: %f\n', sum(spyIntraDayReturn(pomoTopThird)));
legend({'Bottom Third, A/S', 'Middle Third, A/S', 'Top Third, A/S'});
title('POMO and Return');
ylabel('Return');
xlabel('A/S');
[retB, retD, retS] = glmfit(pomoSubmitAcceptRatio, spyIntraDayReturn);
[absretB, absretD, absretS] = glmfit(pomoSubmitAcceptRatio, abs(spyIntraDayReturn));


figure;
hold all;
disp('# Dollar Volume');
scatter(pomoSubmitAcceptRatio(pomoBottomThird), spyDolVol(pomoBottomThird), 'or');
fprintf('Bottom: %f\n', mean(spyDolVol(pomoBottomThird)));
scatter(pomoSubmitAcceptRatio(pomoMiddleThird), spyDolVol(pomoMiddleThird), 'xb');
fprintf('Middle: %f\n', mean(spyDolVol(pomoMiddleThird)));
scatter(pomoSubmitAcceptRatio(pomoTopThird), spyDolVol(pomoTopThird), '.g');
fprintf('Top: %f\n', mean(spyDolVol(pomoTopThird)));
legend({'Bottom Third, A/S', 'Middle Third, A/S', 'Top Third, A/S'});
title('POMO and Dollar Volume');
ylabel('Dollar Volume');
xlabel('A/S');
[dvB, dvD, dvS] = glmfit(pomoSubmitAcceptRatio, spyDolVol);