is it possible to see Time Decay value of an option?

Give me the name of 1 stock you are using this strategy on and the strikes/expiration of the options you are buying/selling at market

 

GLD for example
Call at 147.5 for sep 20th

 

That is a penny wide market. Even if you are paying at market the spread is almost nothing. Are you questioning the fill price you are getting?

 

The markets in GLD are super tight. If you are buying options when the markets are moving fast, you may be getting a sub optimal fill. You are also long vega, so if you are buying when vega is expanding and still holding when the markets are settling then your position could lose value on the volatility contraction.

It's really hard to know without knowing the entire roubd trip details of one of your trades.

 

Yes, fill prices. The way I figure the chains is based on my back testing models and the numbers that deemed to be the best. In back testing models I emulate premium's price based the actual ASK prices for that security that I combine into a matrix and then normalize.
Usually, I am within 1-2% of the price mark, but I believe you are right, because of the volatility factors I can get over 2% over my predicted price quite some times. Gets complicated...

 

A couple of thoughts. First you earlier pointed out that "It does not seem just liner time progression to me." and that is correct, it isn't linear. Just think about the decay in the last 5 minutes of an options life vs 5 minutes of decay on a day 30 days before expiration to intuitively grasp the asymptotic nature of the decay. That aside, you wouldn't expect decay to be linear even between day 30 and day 29 because too many other factors impact volatility over any period, confounding the time decay.
On your second point, you might want to ask yourself if the trigger for your algorithm is set to buy during what are also volatility triggers? For example, if you had a system that implemented an algorithm that bought when a company appeared in the news, you would expect to see the behavior you're seeing of higher than average premiums that not so coincidentally happen right when your algo triggered a buy.