The second graph was just to run a mock simulation to show how a house edge vs no house edge would affect a player's gambling session. You can see the house edge player approaching zero much more rapidly (should be self-explanatory as to why). If you're curious about the simulation parameters, I had a player start with 100 units, bet 1 unit per round on a random color, and have the game result be chosen randomly. You can see the code here if you'd like:
https://pastebin.com/whdJNtG3Well, you seem to be a reasonable person, unlike the other retard but here I doubt if you are missing the point on purpose or what. The problem with the second graph is not that the one who plays a game with HE loses money. The problem is that the one who plays a game with 0 HE has profits. If the time scale is the same as the other line that graph should be equal to the initial balance, 100, at approximately the same time that the one who plays a game with HE loses his money and is left at 0.
In a game with 0 HE the long term tendency should be to stay with the initial balance, and not to have profits, as the blue line shows.