Improving by Statistics - Part Two

If you didn’t see part one, check it out here: https://sc2replaystats.com/blog/archives/1629

 

At the end of the last episode I said I would focus on my three statistically weak points and play the Training Center’s recommended 30 games per week and then report on my progress.

 

What a journey it was! When I made that statement, I had not been playing 1v1 much lately. I had lost my drive to play and found the game often frustrating. I expected those 30 games to be a chore.

 

On the contrary, it reignited my enthusiasm and I had an incredibly fun time. Along the way I had some really memorable games, like when I made a comeback in an epic TvP that ended with mass Battlecruiser vs mass Carrier (he flamed me in Spanish), was called a hacker for doing a 2-1-1 in a TvZ, and had a lot of people calling me some names I shouldn’t write here simply because I went Battlecruisers in TvT.

 

In part one, I was playing defensive styles so I could focus on my macro. Over the course of my 30 games since then I learned this was a bad idea. The longer the game lasts, the harder it is to keep all of your production facilities active and your resource bank low, so the end result was that statistically it appeared my macro was worse. Even though the defensive style helped me macro well in the early-game, it didn’t help me improve the skill of macroing under pressure and still made me look worse on paper.

 

If you’re thinking I went in the opposite direction and started cheesing so it would appear I had improved my macro in this post, I’m happy to tell you that’s not the case. In my 7+ years as a StarCraft II player I’ve rarely cheesed so I don’t really know how. However, I did switch to a more aggressive style in TvP and TvZ, opening 2-1-1  so I could put on pressure early with a lot of Marines and Medivacs while seeing how well I could keep my macro going. I tried to end the game fairly early off of 2-3 base but it rarely worked: I still had plenty of longer macro games. Either way I still used my money a lot better going Bio in those games than I did turtling to Battlecruisers.

 

Notice I didn’t mention TvT. I continued turtling to Battlecruisers in TvT because I find the doom-drops and Tank positioning of TvT too stressful. However, it makes for some less interesting replays and statistics.

So the day of writing this post I decided I wanted to play until I got a fun non-TvT replay to show in the post.

Yep, you guessed it! I got a dozen TvT’s in a row. But in the end, I did finally get a TvZ replay to feature in this post!

Unsurprisingly, there isn’t much change in my statistics over the course of these 30 games. I’ve been playing the game for over 7 years and I’ve been Diamond league for nearly half that time with thousands of games played. After all that time if it were that easy for me to improve my macro in 30 games I would already be Grandmaster.

The statistics don’t tell the whole story, however. The score-screen showed I was was frequently doing better than my average when taken as a whole rather than split up per game stage, and I had a close to 70% win-rate. Overall, I do think focusing on spending helped me overall.

What the statistics also don’t show are skills related to decision making, micro, and army positioning. Is spending money really what’s holding me back, or was it because I sometimes tried to beat Zergs purely with mass speed Banshees or only went mass Battlecruiser in TvT?

I actually think the Sc2ReplayStats Training Center is correct here. I felt nearly all of my loses were due mainly to my poor resource spending not my strategy, positioning, micro, or scouting.

To that end, I would say that the Training Center is a good way to check up on the major things you need to improve on. Just don’t expect to see a big statistical increase in skill across 30 games in one week!

Now on to the replay we’re spotlighting for this week: https://sc2replaystats.com/replay/4500045

You can probably tell the strategy and match-up difference between last week and this week based on the Heat Maps alone:

Last Week (TvT, Defensive)

This Week (TvZ, Aggressive)

I had been going for a quick double Medivac Stim push in TvZ to put pressure on my opponents, but this game my opponent had other ideas.

Lucky my Orbital Command landed just in time, preventing his Baneling bust from being as deadly as it could have been. In fact, I survived with a five worker advantage. Now the only issue is keeping the Zerg from droning too heavily with his three Hatcheries. Luckily, Stim wasn’t canceled and I had my Starport up, so I was still on track for an early Stim-Medivac attack.

I was still a bit behind after the first two Medivacs were driven off, but by continuing to boost over more units, I was able to take the fight I needed.

Despite how cost-effective my trades were, my third was delayed so much that I still was only tied in worker counts, but with MULEs I was starting to pull into a lead.

Unfortunately, when I say my third is late I mean it landed at 10:40 so we’re still neck-in-neck. Bleeding, we both take a moment to defend and get our economy up and running.

Things take a turn for the worse when I’m caught off-guard with un-burrowed Widow Mines.

Luckily, both my macro and my Upgrades are doing well enough the Zerg isn’t quite able to damage my economy significantly. Just after the fight above, my 3-3 finishes a few seconds before his 2-2 finishes.

Like many TvZs, we traded and kept even for many fights, gaining a small edge each time, until suddenly a straw breaks the camels’ back. In this case, in order to trade as well as he did and afford Ultralisks, the Zerg had to cut both upgrades and Anti-Air, leaving his 2-2 Ultralisks to melt to0 Liberators.

Until Next Time!

As always, we love spotlighting community content, so send one of your own replays for us to spotlight next in the comments, or send a suggestion to us on FaceBook or Twitter!

Thanks to everyone who supports our replay spotlights by sharing them with friends, commenting on them, or up-voting them on reddit!

- Sc2replaystats.com staff

 Post written by Justin “TheSkunk” Purdy