Some random thoughts...
1. Taking on the antagonists
There are several trading experts on this board for whom bullying in the cloak of wisdom is their sport. Ad hominem attacks, lying about others and abusive language are their stock and trade. It is useless and dangerous to engage them as they use their anonymity to escalate beyond what most descent people will normally do.
On the other hand, there are those with experience and wisdom who will warn politely of potential danger and may even take the time to answer question.
It's hard to do and I have failed on several occasions but I try to keep in mind the following: "Don't argue with a bully. They will drag you down to their level and beat you with experience."
2. On backtesting period
Backtesting can be deceptive. The backtest period should be long enough to include all market conditions you are measuring at least twice over. More on that later.
3. On backtesting data
I have yet to see a data set that was error free and I have been backtesting since before you were even in liquid form. I am a software engineer, so most of my backtesting was my own code. The first task was to write an expert system to test the data set and through out the bad data.
4. On backtesting algorithms.
There is a maxim in software engineering: "In programming you are always off by 1." I'm embarrassed to tell you how many times I thought I had a great strategy only to find that I had accidentally looked into the future. Some of my bugs were quite subtle and hard to detect.
5. Curve fitting
One of the most common mistakes is to optimize your strategy to the point that it gets wonderful results on your training data set. This is not your goal. Your goal is to discover some general truth that will also give good results outside of the test data set.
6. Backtesting technique.
I go to extraordinary measures to eliminate "curve fitting," but then I'm a fanatic on the subject. At a minimum I recommend dividing your historical data into two data sets; one for training and one for validation. Only if the results are similar do you have any confidence that you have generalized your strategy to survive all the market conditions it uses.
7. Backtesting fanaticism.
To give you a small idea the extent fanatics will go, permit me to give you some idea of what I do. First, each run against my test strategy goes against a randomly selected subset of my data. Each data set is generally within 40% to 60% of the entire set. I tweak my strategy until I get "similar results" between the training data and the validation data.
Now I do thousands of runs (Monte Carlo analysis) and get some statistics that measure consistency within each data set type (training and validation) and between data set types. If your set of training runs show a wide variance, I would worry.
8. Manufacturing analogy.
From working for a mechanical components manufacturing company I learned the four steps from idea to production and use their equivalent in qualifying a strategy for serious money trading.
Computer simulation -- any time we did a new design we first designed it within design programs where we could perform stress and performance prediction on the part. --
This is backtesting in the trading world.
Prototype testing -- we built prototype parts to find out how they interfaced with the real world and to validate that our simulations represented reality. --
This is SIM or paper trading in the trading world. You need to find out the market rhythm of the trading vehicle and what is required of the trader (or robot) to execute successfully. I never fail to learn something I did not expect when looking a trading vehicle that is new to me. And I have never performed as well in SIM trading as I did in backtesting.
Preproduction -- once a part has qualified with prototype testing, we do a preprpduction run. Here you get the real world of manufacturing and you find out if you can hold the tolerances and keep the production speed up high enough to be profitable --
This is small money trading. Here you encounter real world fills rather than the fills you project in SIM trading. You get to verify many of your assumptions you made in backtesting and SIM trading.
Production -- ready to rock and roll!! Here we crank up the line, but put in place monitors (guaging) to quickly spot failures in the system, both anticipated and unanticipated. --
Serious money trading In trading you have similar concerns. You now have a trading system; a strategy and the trader (you). A good strategy traded by a failing trader will still lose money. The trick is to put measurements in place to spot trader failure (and fix it) vs. strategy failure. I use a trading journal for this purpose.
_____________________________________________________
You're are young, inquisitive and willing to try stuff. I'm excited for you. Hopefully, I have given you some insights into the tools that you may find helpful.
A book I would recommend is
The Evaluation and Optimization of Trading Strategies It's expensive, but your library may be able to borrow one via inter-library loan.
. My father always tells me its the haters that fuel your fire to succeed. You arguing with them doesn't do anything, just prove them wrong without words.