Backtest of Forecasting

I'd like to backtest my forecasting. I used the template from "forecsting the US economy". Instead of dates I used observations. Since I'm still a beginner with Matlab, I'd appreciate if someone helped me. I don't really get the last part of the code and don't know how to convert it. I'm sorry.

The data I used is stored in a double array, containing 5372 observations.

%% Backtest

syy = Return(3000);                 % start year for backtest
eyy = Return(5372);                 % end year for backtest

Horizon = 1;             % number of quarters for forecast horizon

[T, n] = size(Y);

FError = NaN(eyy - syy + 1, n);
FDates = zeros(eyy - syy + 1, 1);

for yy = syy:eyy
  
  StartDate = Return(3000);
  EndDate = Return(5372);

% Locate indexes in data for specified date range

        iStart = find(StartDate <= dates,1);
  if iStart < 1
    iStart = 1;
  end
  iEnd = find(EndDate <= dates,1);
  if iEnd > numel(dates)
    iEnd = numel(dates);
  end

        if iStart > 1
    Y0 = Y(1:iStart-1,:);
  else
    Y0 = [];
  end
  Y1 = Y(iStart:iEnd,:);
  iY1 = iY(iStart:iEnd,:);

% Set up model and estimate coefficients

    spec = garchset('VarianceModel','GJR',...
                'R',1,'M',1,'P',1,'Q',1);

    spec = garchset(spec,'Display','off','Distribution','T');

    [coeff,errors,LLF,eFit,sFit] = garchfit(spec,Return);
  
  % Do forecasts

horizon = 1 [sigmaForecast,meanForecast,sigmaTotal,meanRMSE] = ... garchpred(coeff,Return,horizon);

% Assess Forecast Quality

Ow = max(0,min(Horizon,(size(Y,1) - iEnd)));  

if Ow >= Horizon
    h = Horizon;
  else
    h = [];
  end

FDates(yy-syy+1) = yy;
  if ~isempty(h)
Yerr = iY(iEnd+1:iEnd+Ow,:) - eFY(1:Ow,:);

Ym2 = Yerr(1:h,:) .^ 2;
Yrmse = sqrt(mean(Ym2,1));
    
fprintf('%12s',datestr(EndDate,1));
for i = 1:n
fprintf('  %7.2f',100*Yrmse(i));
end
FError(yy-syy+1,:) = 100*Yrmse';
fprintf('\n');
  end
end

% Accuracy

mFError = NaN(size(FError));
sFError = NaN(size(FError));
for i = 1:n
mFError(:,i) = nanmean(FError(:,i));
sFError(:,i) = nanstd(FError(:,i));
end

for i = 1:n
subplot(ceil(n/2),2,i,'align');
plot(FDates,FError(:,i));
hold on
plot(FDates,mFError(:,i),'g');
plot(FDates,[mFError(:,i) - sFError(:,i),mFError(:,i) + sFError(:,i)],':r');
dateaxis('x',1);
if i == 1000
 title(['\bfForecast Accuracy for ' sprintf('%g',Horizon) '-Year Horizon']);
end
end