From what i see, different PC gamers want to see different things. i have to agree with a need for newer titles. But then there are a lot of (sensible) gamers who only pick up games when they hit the bargain - after they are fully patched and the bugs (and the DRM) are completely worked out. Now we are talking a LOT of people buying new games for themselves when the games are 6 months to two years old. Also, people with midrange PCs wisely do this.

As for me as a reviewer, i feel the need to actually PLAY a game before i can include it. Some of them i don't like and i will never finish. Others i am getting late (like Crysis 2 when it is fully DX11). Others require an understanding of the relationship of the benchmark to the game and how to relate it to the reader. Some require creation of a time demo or a Fraps run (which is hard to reproduce multiple times).

i am going to add Shogun 2 to my benching suite but you will never catch me playing SC2 nor benching it (the results are too variable imo across the testing sites). i have Bulletstorm and will attempt to bench it; probably not unless a cutscene is generated ingame under the right circumstances - but that takes research and trial and error.

Finally, i don't really like to go back and play old games unless i am playing them in 3-panel Surround or Eyefinity - or perhaps in 3D .. then it is great to revisit old games and know how they perform at 5760x1080 - perhaps in 3D at 120Hz

I agree older games in 3d and surround is a good idea also.

Rollo?, I've been called a lot of things but rollo?

OK say I'm a gamer and I have a gtx460 and I just played through Crysis 2 at high settings at above 40 fps minimum @ 1080p.

Now I want to upgrade my video card so I go check out some reivews. Why should I care that the new gtx 680 will run Crysis 2 at 100 fps the following year. I much rather find out that my gtx460 will not cut it for Crysis 2 expansion pack 1 direct x 11 super pack and I need to upgrade to the new gtx680.

I'm not saying its bad to test some older games ,but I still see review sites using Left for Dead 1 with a 6970 @ 250 fps. For gods sake the 2nd Left for Dead is allready getting old.

As gaming has become more and more mainstream, we see huge marketing dollars and ultra hype on blockbuster titles. The media industry has found a dead lock grip and structured mainstream gaming so well the billions stack higher and higher each year. These mega game releases are a big deal, the launches are bringing in record revenues which is very impressive. Most of the time a few months is all one needs to declare a game a success or a flop. But not all gaming platforms follow the same pattern, PC gaming differs greatly from the console world in a very important way.

Console releases that dont sell well out the gate usually end up flopping. PC gaming isnt like this. The industry, it seems. is more aware of this now. When Crysis first came out it didnt do well at all! There was a huge fear that it wasnt gonna get close to breaking even. But steady the sales came, instead of decreasing over time the the volume went up. Now this many years later we still see crysis everywhere. It didnt fail, it was a huge success. The PC market is vastly different in the fact that PC gamers tend to faithfully buy yesterdays hit titles. Every gamer is different, but PC is a platform that has sales build over long periods of time.

As far as benchmarking, i must say that its not useful to have only dated games. But just as well, neither is it good to only have the few hits out that month. The larger the sample, the more valuable the data is. You have to understand that while the new games may vary greatly, the more data sampled the better. It becomes less and less skewed. A title from 2yrs ago does nothing to hurt, it only gives more data. If its not a game you care about, why focus on it? The real truth is the larger the sample of data, the better the results are. The more clear the picture becomes. Titles come and go, the engines change over time. Jusyt doing current hits is so limiting. In a month or to things can be very different as the next few hits come out. The bigger the review sample from the start, the more accurate the average will be. This average will continue on as future games will continue to be vary wildly.

I am all for the data. The more there is, it only gets better and better. I hope to see apoppin stay true to the large benchmark samples. this is what makes his reviews so important to me. I think he has a great mix. If all you want to see is the new title results, the just focus on them, as the data will still be there. And the results can be yours to interpret. Of course he will need to update the titles as time goes by. Why ask to not include data? It can only be useful for a more complete picture. It cant hurt anything as it only makes the results more and more complete.

there is very little performance difference using any of the recent versions of the Source engine. Running L4D at 2560x1600 or 5760x1080 at 8xMSAA is not exactly "light" on a PC - nor is is much less demanding than running L4D2 at those same settings.

It's Valve's fault.... WtH is HL 3 ?